Copyright 2008, Jared Michael Hansen

112
Retail Price Promotions and Retailer Financial Performance: The Roles of Expandable Purchases, Inventory Turnover Acceleration, and Competitive Intensity by Jared M. Hansen, B.S. in Engineering, MBA A Dissertation In BUSINESS ADMINISTRATION - MARKETING Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY Approved Shelby D. Hunt Chairman of the Committee James B. Wilcox Dale F. Duhan Steve Buchheit Fred Hartmesiter Dean of the Graduate School December, 2007

Transcript of Copyright 2008, Jared Michael Hansen

Page 1: Copyright 2008, Jared Michael Hansen

Retail Price Promotions and Retailer Financial Performance: The Roles of Expandable Purchases, Inventory Turnover Acceleration, and

Competitive Intensity

by

Jared M. Hansen, B.S. in Engineering,

MBA

A Dissertation

In

BUSINESS ADMINISTRATION - MARKETING

Submitted to the Graduate Faculty of Texas Tech University in

Partial Fulfillment of the Requirements for

the Degree of

DOCTOR OF PHILOSOPHY

Approved

Shelby D. Hunt Chairman of the Committee

James B. Wilcox

Dale F. Duhan

Steve Buchheit

Fred Hartmesiter Dean of the Graduate School

December, 2007

Page 2: Copyright 2008, Jared Michael Hansen

Copyright 2008, Jared Michael Hansen

Page 3: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

ACKNOWLEDGMENTS

I extend appreciation to all those who have assisted me as I completed this

dissertation, including the professors and doctoral students of the Texas Tech

University marketing department who have been willing, insightful sounding boards

for the refinement of many of the ideas presented in this dissertation. In particular, I

thank Dr. Shelby Hunt for being a mentor and friend, in addition to chairman of my

dissertation, and for the learning I have gained through daily interactions I have been

privileged to have with him during the last two and a half years that I have been his

teaching assistant at Texas Tech. I also thank Dr. Jim Wilcox for his Aristotelian

council on how I might improve measurement in the dissertation. Likewise, I

appreciate the passion that Dr. Dale F. Duhan exhibited on the subject of market

basket analysis. Our many conversations on the subject have been most helpful. I also

thank Dr. Steve Buchheit for helping me attempt to bridge research from accounting,

finance, and marketing.

I give a very deep thanks to the several unnamed friends and associates in

business practice who have made this research possible by providing data. Without the

contributions of these executives and senior managers, the hypotheses could not ever

be explored in their current forms.

I also thank several individuals who have steered me to where I am today.

First, I thank my parents for teaching me to love learning. I thank Ms. Evelynn

Bassutt and Dr. Jeff Campbell who were great examples of teaching in excellence and

motivated me to excel in my studies. I thank Chris Gould and Chris Black who hired

me to be a corporate buyer at Wal-Mart, and Alan Epler, Tom Daugherty, Terry Clark,

and others who were mentors and friends while I was in Bentonville. In all, my

teaching and research compliment each other, and both draw heavily from my industry

experiences. Last, and most important, I thank my wife and best friend Tracy who,

along with my children, made many sacrifices so that I could complete this

dissertation.

ii

Page 4: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

TABLE OF CONTENTS

ACKNOWLEDGEMENTS .................................................................................... ii

ABSTRACT ........................................................................................................... iv

LIST OF TABLES .................................................................................................. v

LIST OF FIGURES ............................................................................................... vi

CHAPTER 1. INTRODUCTION ........................................................................... 1

CHAPTER 2. A REVIEW OF THE LITERATURE ........................................... 26

Overview ............................................................................................................... 26

Price Promotions and Definition ........................................................................... 26

Price Promotions and Expandable Purchasing ...................................................... 30

Price Promotions and Return Calculation ............................................................. 41

Moderations of the Promotion-Performance Relationships .................................. 50

CHAPTER 3. RESEARCH DESIGN ................................................................... 57

Samples ................................................................................................................. 57

Data Collection ..................................................................................................... 60

Measurement of Variables .................................................................................... 62

CHAPTER 4. DATA ANALYSIS ........................................................................ 69

CHAPTER 5. DISCUSSION AND CONCLUSION ........................................... 77

REFERENCES ...................................................................................................... 96

APPENDIX ......................................................................................................... 105

iii

Page 5: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

ABSTRACT This dissertation investigates the effects of pure, retail price promotions on

product performance, market basket performance, and shareholder investment in the

retailer’s product-level inventory. By pure, I refer to the absence of any additional

displays, features, coupons, advertisements, or other intentional activities by the

retailer that could confound the effects of the price promotions, themselves. The

underlying research question of this study is: What is the impact of retail price

promotions on retailer performance? Specifically, I focus on price promotions and

two measures of performance: (1) the retailer’s total market basket performance (i.e.,

market basket sales, profitability, and inventory turnover), and (2) the price-promoted

products’ financial return on shareholder investment in product-level inventory.

Furthering the goal of understanding the effect of promotional activities on

retailer performance, this study is the first study to (1) measure, instead of estimate,

the effect of individual price promotions on the actual, total-market-basket

performance, and (2) quantify the impact of retailer promotional activity on the

retailer’s shareholders by introducing and using a new metric, the gross margin return

on shareholder investment (or GMROSI) in the retailer’s product-level inventory. The

study also examines several moderating variables, including competitive intensity.

The sample for this dissertation consists of weekly, point-of-sale, stock-

keeping unit, store scanner data for manufacturer-branded products collected in all of

the approximately 450 stores of an every-day-low-price retailer operating in the

northeastern and midwestern United States.

iv

Page 6: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

LIST OF TABLES 1. Literature Related to the Effects of Price Promotions on Market Basket

Performance .................................................................................................. 83

2. The Effects of Pure Price Promotions on Product Performance ................... 84

3. Effects of Pure Price Promotions on Market Basket Performance ............... 85

4. Standardized Ridge Regression Coefficients-GMROII Response Surface .. 86

5. The Effects of Pure Price Promotions on GMROII and GMROSI ............... 87

6. The Moderating Effects on H3-H7 ............................................................... 88

v

Page 7: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

vi

LIST OF FIGURES

1. A Causal Model of The Effects of Retail Price Promotions ..................... 89

2. Retailers’ GMROII Response Surface Map ............................................. 90

3. Retailers’ GMROII Response Surface Map with Promotion Data Overlay ............................................................................................. 91

4.a. Response Surface Map of Sample A Marketi Basket

Profitability ............................................................................................... 92

4.b. Response Surface Map of Sample B Marketi Basket

Profitability ............................................................................................... 93

4.c. Response Surface Map of Sample C Marketi Basket

Profitability ............................................................................................... 94

4.d. Response Surface Map of Sample D Marketi Basket

Profitability ............................................................................................... 95

Page 8: Copyright 2008, Jared Michael Hansen

CHAPTER 1

INTRODUCTION

The concept of retailer price promotions has been an important topic in

marketing practice and academe for at least a century. A major impetus for retailer

price promotions was an early US Supreme Court ruling that manufacturers cannot set

retail minimum markup prices (Miles v. Park and Son 1911). A more recent impetus

has been the ongoing shift in relational influence from manufacturers to retailers in

supply chains (e.g., Fishman 2006; Hansen 2007; Mulhern and Leone 1991; Useem,

2003). These forces have afforded retailers the opportunity to control their market

offering retail prices. As Walters (1991) points out, “the primary purpose of retail

prices is to increase retailer sales, and, in turn, retailer profitability.” As a

consequence, retail price promotions have become an “important tool for the modern

marketing managers in stimulating sales” (Goodman and Moody 1970, p. 31).

Retail price promotions can affect the promoted product’s inventory turnover,

as well as influence the merchandise category level performance (e.g., sales,

markdowns, receipts, turns, gross profit, and gross margin return on inventory

investment or “GMROII”) (Hansen, Raut, and Swami 2006, forthcoming). For

example, some price-promoted products may increase category sales but decrease the

category-level (average) initial margin. Other product price promotions may increase

the category initial margin percentage, but retard category turnover. Thus, successful

retailers must adjust to their changing environments, or change them, by deciding (1)

how much to price promote (i.e., how deep to discount), (2) which products to price

Page 9: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

promote, (3) where to price promote (i.e., in which stores), and (4) when to price

promote.

The purpose of this research is to increase our understanding of the impact of

pure, retailer price promotions (hereafter, price promotions). By pure, I refer to the

absence of any additional displays, features, coupons, advertisements, or other

intentional activities by the retailer that could confound the effects of price promotions

(see Van Heede, Leeflang, and Wittink 2004, for additional discussion on “pure”

promotions). Thus, they are distinct from products that are retailer-advertised but not

retailer-price-discounted (e.g., to create awareness of the product availability at the

store). They are also distinct from products that are retailer-advertised and retailer-

priced-discounted (e.g., to increase customers visiting the store). (Pure) price

promotions, in contrast, are used by retailers in an attempt to increase purchasing

volume across customers and over time (e.g., Blattberg and Wisniewski 1987;

Duncan, Hollander, and Savitt 1983; Kuehn and Rohloff 1967; Mason and Mayer

1984; Moriarty 1985; Woodside and Waddle 1975), and are often the most frequent

type of promotion used by retailers (e.g., Gedenk, Neslin, and Ailawadi 2006),

especially for every-day-low-price retailers such as Wal-Mart that dominant the retail

industry (Heller 2001).1

1 Consider, for instance, that while Wal-Mart reported 718 price promotions (i.e., “rollbacks”) in May 2007 (Wal-Mart 2007), only 31 of the price promotion were included in their printed monthly advertisement circular—in addition to the 70 items advertised in the same circular that were not being price promoted. Combining the promotional counts, 9% of promotions were advertised, but not price discounted, 4% were advertised and price discounted, and 87% were price discounted and not advertised.

2

Page 10: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

The underlying research question of this study is: What is the impact of

retailer price promotions on retailer performance? Specifically, I focus on price

promotions and two measures of retailer performance: (1) the retailer’s total market

basket performance (i.e., market basket sales, profitability, and inventory turnover),

and (2) the price-promoted products’ financial return on shareholder (inventory)

investment. A process model indicating the hypotheses to be explored is presented in

Figure 1. The uniqueness of the research sample design is presented in Table 1.

This dissertation proceeds as follows. First, hypotheses are presented regarding

the effect of a product’s price promotions on that product’s sales, inventory turnover,

and profitability. These hypotheses are intended to ensure that the samples used in this

research are consistent with prior retail price promotion research samples. Therefore,

these hypotheses replicate previous research on price promotions. Second, a rationale

of price promotions is presented that is consistent with a theory of (what I label)

expandable purchases. This view of purchasing and consumption is used to develop

hypotheses related to the first investigated measure of retailer profitability, that is, the

impact of price promotions on the total market basket. Third, the chapter presents an

overview of a new financial performance metric for evaluating price promotions. This

new metric incorporates the effects price promotions have on the shareholder return on

investment. Fourth, the chapter presents an initial discussion on how competitive

intensity affects the hypotheses on product performance, market basket performance,

and shareholder return on investment. Last, the chapter outlines the research design of

the dissertation.

3

Page 11: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Replication Hypotheses

Prior literature has found that retail price promotions have a positive effect on

product sales volume (i.e., units sold) and dollars (i.e., net sale dollars) during the

promotional period (e.g., Ailawadi and Neslin 1998; Van Heerde, Gupta, and Wittink

2003). Related, it seems intuitive to hypothesize that the increase in sales volume also

typically accelerates the inventory turnover rate. However, when retail price

promotions increase sales volume, how often (or when) is the increase sufficient to

improve profitability (i.e., initial gross margin dollars)? For instance, a price

promotion of Oreo’s could result in $50 sales growth, a $10 gross margin loss, and a

one-fold increase in inventory turnover. Or, it could result in a $100 sales growth, a

$20 gross margin growth, and a two-fold increase in inventory turnover. Given the

substantial, previous research on the effects of price promotions on sales volume and,

in contrast, the paucity of previous research on the effects of price promotions on

profitability, I hypothesize (and ask):

H1: Price promotions of a product will have a positive effect on that product’s sales (at the product-store level).

H2: Price promotions of a product will have a positive effect on that product’s

inventory turnover (at the product-store level).

RQ1: What is the effect of price promotions of a product on that product’s profitability (at the product-store level)?

Expandable Purchases and Market Basket Performance

There is a growing stream of literature that finds that retailer price promotions

can not only increase the purchase of promoted product(s), but also can result in

changes in the purchases of nonpromoted, substitutable and complementary products

4

Page 12: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

(e.g., Ailawadi et al. 2006; Ailawadi and Neslin 1998; Chandon and Wansink 2002,

Chintagunta and Haldar 1998; Janakiraman, Meyer, Morales 2006; Mulhern and

Leone 1991; Walters and MacKenzie 1988; Walters 1991). Consequently, there are

calls for price-promotion research to move from a product or brand focus to a category

management level of analysis to account for customer product choice substitutes (e.g.,

from focusing on Oreos to focusing on cookies; see Nijs et al. 2001). Even further,

some researchers argue for price-promotion research to shift from category

management to investigating total store purchases to account for customer product

choice complements (e.g., Bucklin and Gupta 1999; Manchanda, Ansari, and Gupta

1999; Shocker, Bayus, and Kim 2004). In particular, two recent articles have provided

initial evidence that price promotions can increase total shopping cart (i.e., store-level,

market basket) purchases per shopping visit (Ailawadi et al 2006; Janakiraman,

Meyer, Morales 2006).

The proposed research advances the literature on the price promotion-retailer

performance relationship. Specifically, I develop and test hypotheses relating the price

promotions of a product to changes in the customer’s market basket. Further, I

quantify the impact of retailer promotional activity on the retailer’s shareholders by

introducing and discussing a new financial performance metric. Moreover, I

investigate the role of competitive intensity as a moderator of the relationship between

product price promotions and retailer profitability, identifying when promotions work

well for retailers (i.e., customer, demographic-based, market segments).

5

Page 13: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

As indicated in the research question, one contribution of this research is to

provide the first documented analysis of the impact of pure price promotions of

individual products on the total purchase of the customer per shopping visit (hereafter,

market basket) using store-level scanner data. According to Hansen, Raut and Sumit

(forthcoming, p.1), “research is needed that investigates the effect of price promotions

…on market basket performance.” While some research has begun to explore the

potential, particular product, complement (e.g., inter-category performance) and

product substitute (e.g., intra-category performance) effects in market baskets (e.g.,

Shocker, Bayus, Kim 2004), no study has looked at the impact of individual product

pure price promotions on total-market-basket sales, profitability, and inventory

turnover. See Table 1. The absence of research on total-market-basket effects could be

because of the difficulty of acquiring data, which is usually available only through

proprietary, store-level scanner data.

For example, the recent work by Ailawadi et al. (2006, p.526) is “the first to

estimate halo [i.e., market basket expansion] rates.” However, they use a regression

based estimate of the quantity of items in the market basket based on a sample of store

loyalty card information because they do not have access to market basket information

for the store-level scanner data. They then estimate the market basket sales dollars

and gross profit dollar by multiplying the estimated change in units by the total store

sales and profits averages. There are several limitations to such store-level estimations,

one of which is the inability to calculate the market basket metrics for each individual

promotion (and individual promotions are, indeed, the specified unit level of analysis).

6

Page 14: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Therefore, Ailawadi et al (2006, p. 531; emphasis added) call for future researchers to

“validate our work, find new ways of obtaining more disaggregate estimates of the

halo effect, and study how and why halo effects vary across categories and retail

formats.”

Furthering the goal of understanding the effect of promotional activities on

retailer performance, this study is the first study to measure, instead of estimate, the

effect of individual pure price promotions on the actual, total-market-basket

performance. See Table 1. Synthesizing the findings of prior research on the subject

of market basket (e.g., Ailawadi et al 2006; Janakiraman, Meyer, and Morales 2006;

Mulhern and Padgett 1995), this research proposes that purchase and consumption are

expandable. One example is the “halo effect” proposed by Ailawadi et al. (2006).

Another example is the “spillover effect” proposed by Janakiraman, Meyer, and

Morales (2006). Other examples are found in the works of Chandon, Wansink,

Laurent (2000), who identify six major multiple consumer benefits from sales

promotions: opportunities for value expression, entertainment, exploration, savings,

higher product quality, improved shopping convenience. Also, McCracken (1999)

presents several other potential reasons for consumption, including identity, emotional

fulfillment, and a consistent product constellation. This dissertation proposes that

(what I refer to as) expandable purchasing can occur as a result of any of these, or

perhaps others, individually or in combination.

The word “expandable” indicates that, contrary to the commonly-accepted

“market pie” metaphor in marketing (e.g., Carson et al. 1999; Chakravarti and

7

Page 15: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Janiszewski 2004; Day and Montgomery 1999; Jap 1999; Nason 2006; Rokkan,

Heide, and Wathne 2003; Weitz and Bradford 1999), consumption is not like a static

“market pie” to be divided (i.e., maintained). This usage of the pie metaphor is an

adaptation of the neoclassical economic pie metaphor. Speaking on the original usage

of the metaphor in macroeconomics, Swinnerton (1997, p. 75) states, “To develop a

better intuition for the definitions of efficiency and their equivalence, it is useful to

think of a pie as a metaphor for the output of the economy.” In the pie metaphor in

marketing research, however, researchers have equated pie size with “primary

demand” (instead of labor allocation) and pie distribution with “market share” (instead

of wealth distribution). The pie metaphor, as currently applied to purchasing-related

research, implies that all consumers go to a store with a budget and they spend the

budget. This adaptation of the metaphor has resulted in interpretations of price

promotions as zero-sum games of product switching (e.g., Dodson, Tybout, and

Sternthal 1978).

The growing literature converging on a theory of expandable purchasing does

not support this metaphor or its resulting interpretation of price promotion research.

Instead, this growing stream of research is more consistent with a balloon metaphor,

where purchasing can be grown similar to the inflating of a balloon (Hansen and

McGinty, 2007). Rather than assuming that all consumers go to a store with a budget

and spend the budget limit, this approach assumes that consumers may visit with a

budget in mind, but retain the flexibility to adjust spending. For instance, a consumer

may plan on spending 100 dollars as he has done in the past, but decides to spend

8

Page 16: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

more because of the effect of the pure price promotion. Similar to the balloon analogy,

a shallow discount will not result in much of an effect on overall purchase behavior.

However, the effect grows nonlinearly with the depth of the price promotion. That is,

as the size of the price promotion for a product increases, there is expanded purchasing

of other products. These products may be complements or may be completely

unrelated. Indeed, according to Walters and MacKenzie (1988, p. 54), “Retailers and

marketing researchers generally believe one of the primary benefits of price

promotions is that they stimulate sales not only for the lower price, lower margin

promoted items but also for higher margin goods that are not being promoted.”

The motivation for expanded purchasing may be economic utility

maximization, hedonic enjoyment, a reward to the retailer, or a combination of these

and other consumer motivations. I propose that there is a point (though it might be

difficult to locate precisely) where a very deep price promotion might result in

negative reaction by consumers who begin to doubt the quality of the product given

the large price reduction (similar to the balloon popping). See Kirmani and Rao (2000)

for discussion on signaling unobservable product quality. Most retailers, however, do

not engage in this kind of price promotion activity, with the exception of occasional

loss leaders (e.g., milk, bread, crayons, beer). Rather, the relevant range of the type of

retail price promotions being investigated (i.e., context) is expected to be in the “front

half,” or positive slope area, of the proposed balloon effect. Based on the preceding

discussion, I hypothesize:

H3: Price promotions of a product will have positive effects on market basket sales.

9

Page 17: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

H4: Price promotions of a product will have positive effects on market basket

item count (i.e., number of products in the market basket). H5: Price promotions of a product will have positive effects on market basket

profitability.

Shareholder Return on Investment Performance

Another contribution of this research is that it investigates the effect of price

promotions on shareholder investment. There have been calls (e.g., Lehmann 2004;

Rust et al. 2004) for marketing research to build bridges between marketing activities

and financial (e.g., shareholder) outcomes. Further, recent accounting research

recognizes and calls for research that addresses the difference between ROI and the

real economic profitability of an organization (see Rajan, Reichelstein, and Soliman,

forthcoming). This research will be the first study, based on a review of the literature

and discussion with several thought leaders in the area, to quantify the impact of

retailer promotional activity on the retailer’s shareholders by introducing and using a

new metric.

Early works on valuation often used return on investment (ROI).

Unfortunately, the “investment” in ROI varies by research study. The formula

variation (in studies purporting to use ROI) has resulted in several noted potential

problems in interpretation. Most often ROI appears to become return on sales—

which should be identified as ROS, not ROI. ROS is calculated as:

(1) ROS = Net Income (Before Interest & Tax) ÷ Sales Dollars

10

Page 18: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Practitioners are usually more interested in measuring performance within their

organization (e.g., particular product SKU price promotions) versus measuring

performance across organizations. Attempting to apply ROI (or ROS) to product (or

even category) merchandise decisions is, as indicated by Sweeney (1973, p. 61),

“fraught with cost measurement and allocation problems.” Further, price promotions,

by mathematical rule, decrease the unit margin in order to stimulate sales growth. In

the context of ROS, the growth in the numerator (i.e., net income) is always less than

the growth in the denominator. Take, for example, a box of Tide detergent that is

regularly prices at $6, which generates $2 in net income and a ROS of 2/6=33%. The

Tide is price promoted at $5 per box, which decreases net income to $1 and results in

a ROS of 1/5=20%. While there may be a ten-fold increase in sales volume, increasing

both the sales dollars and income dollars, the ROS ratio decreases. Thus, neither ROI

nor ROS is well equipped to measure product-level, operational performance.

However, both the allocation and dollars versus percentage problems can be overcome

when “the performance measure is used exclusively for planning and controlling

merchandising inventory investment” (Sweeney 1973, p. 61).

Consequently, a few academic researchers and many practitioners have

adopted a more specific measure of return, the gross margin return on inventory

investment (hereafter, GMROII) to plan and control (e.g., Ahern and Romano 1979;

Dunne and Lusch 2005; Kravitz 1977; Leeds 1976; Sweeney 1973; Tolle 1976;

Warrington 1982). GMROII is calculated as:

(2) GMROII ( % ) = Gross Margin (%) ÷ [1 - Gross Margin (%)] x Inventory Turnover

11

Page 19: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

One advantage of GMROII is that it does not require any approximations from

aggregated activities such as labor costs. Thus, it can be applied at the product,

category, or overall retailer level of analysis. Another advantage of GMROII (over

ROS) is that GMROII accounts for inventory turnover. Also, while other return

metrics (e.g., ROI, ROS) do involve a time component (i.e., it is the return on

investment or sales for a given period of time), GMROII emphasizes the importance

of time. That is, it shows how time can provide a strategic advantage/disadvantage as

an organization manages its inventory (i.e., cost of goods sold) better/worse than its

competitors. Despite these advantages, there is a dearth of academic literature that

uses GMROII. Perhaps one reason why few academic researchers use GMROII (in

contrast to many practitioners) is access to data on both profitability and inventory

turnover rates. One result is that the effect of the price promotions of a product on

GMROII has not been established in the literature—an important gap that needs

addressing. This effect could be either positive or negative, depending on how the

gross margin percent and the inventory turnover are conjointly affected by the price

promotions.

For example, a price promotion of widgets might result in both a ten percent

decrease in gross margin percent (i.e., from 50 percent to 40 percent) and a three fold

increase in inventory turnover (i.e., from two to six turns annually). Or, a price

promotion of the same widgets might result in both a five percent decrease in gross

margin percent (i.e., from 50 percent to 45 percent) and a one fold increase in

inventory turnover (i.e., from two to three turns annually). The first scenario (smaller

12

Page 20: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

gross margin, larger turnover) results in a GMROII of 400 percent return, whereas the

second scenario (larger gross margin, smaller turnover) results in a GMROII of 245

percent return. In this example, the price promotion that resulted in less of a margin

decrease did not result in a higher GMROII (because of the relative inventory turnover

impact). Thus, in hypothesizing the impact of price promotions on GMROII, a

researcher is implicitly hypothesizing about the relative change in gross margin versus

inventory turnover rate. The neoclassical economic “pie” metaphor (presented earlier

in this chapter) maintains that the inventory rate is maintained over time or across

products (i.e., substitution effects). In contrast, a theory of expandable purchasing

suggests that the inventory turnover rate would be accelerated due to the product’s

price promotion. Based on the preceding discussion, and consistent with the growing

literature supporting the phenomenon that I refer to as expandable purchasing, I

hypothesize:

H6: Price promotions of products, in a majority of cases, will have positive effects on promoted products’ GMROII.

As a derivation of ROI, though, GMROII does not consider the financing of

the cost component, and thus retains the inability to document the impact of price-

promotions on an organization’s owners. The cost component consists of both

financing and operating responsibilities, and operational metrics such as ROI and

GMROII follow managerial accounting guidelines that have traditionally held that:

Managers have both financing and operating responsibilities. Financing responsibilities relate to how one obtains the funds needed to provide for the assets in an organization. Operating responsibilities related to how one uses the assets once they have been obtained. Both are vital to a well-managed firm.

13

Page 21: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

However, care must be taken not to confuse or mix the two when assessing the performance of a manager. (Garrison and Noreen 2003, p.773)

The foundational logic of this research agrees with the premises of these

quoted guidelines. That is, there are financing and operating responsibilities (i.e.,

premise 1) and both responsibilities are vital to the organization (i.e., premise 2).

However, it is proposed that the inventory turnover rate does have an effect on the

accounts payables. That is, the inventory turnover rate can change how one obtains the

funds (e.g., what is funded internally, or funded externally, or, even, never requires

funding at all). In contrast to ROI, return on equity (ROE) considers the financial

leveraging of funding (see Block and Hirt 2000, p. 56):

(3) Return on equity = Return on assets (investment) ÷ (1-Debt/Assets)

However, the ROE metric requires assumptions about dividing the costs of

labor, space, and other activities to arrive at net income that are not easily allocated to

product level analysis. At the same time, the cost of financing versus investing is

being debated (see Baker, Ruback, and Wurgler 2007). Thus, an alternative metric in

use is proposed here that is similar to the Du Pont model of ROE (see Garrison and

Noreen 2003, p. 70), and can be used as a framework for discussing promotional

impact on the organization. This metric replaces net income with gross margin (i.e.,

the purchased cost of goods, prior to allocating building and other costs) and, in

accounting for the change in the financing plan, also incorporates the internal rate of

return on the unused assets. Thus, the metric becomes the gross margin return on

shareholder investment (hereafter, GMROSI) of the product.

14

Page 22: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

In relation to GMROII, the GMROSI metric replaces the “inventory”

investment with “shareholder” investment by accounting for the change in the cost due

to changes in account payables that are, in turn, due to changes in the contracted terms

of payment (e.g., 2-10-net 60) between the retailer and the manufacturer of the

product, and the actual (i.e., “positive,” not “normative”) internal rate of return in the

organization. By doing so, this planning and controlling metric accounts for the

relative return on each shareholder dollar invested in the organization (i.e., capital cash

flow productivity). See Appendix 1 for an example. Also, while GMROSI provides an

operational accounting of shareholder inventory investment productivity, it should not

be confused with shareholder market return. (Shareholder market return--as

traditionally used in accounting and finance--accounts for the variance in stock price.)

The relevant formulas for GMROSI computation are:

(4) GMROSI ( % ) = Adjusted Gross Margin (%) ÷ [1 – Adjusted Gross Margin (%)] x Inventory Turnover, where

(5) Adjusted Gross Margin = (Sales $ - Adjusted Cost $) ÷ Sales $, where

(6) Adjusted Cost $ = (Cost Schedule * Cost%) - ( IRR * Cost Schedule

* [1-Cost%] )

Taking a shareholder perspective, it is proposed that many promotions that

decrease the percent profitability of the product for the retailer may, in fact, increase

the percent profitability of the product for the shareholder. The argument here runs

contrary to the findings of Srinivasan et al. (2004, p. 617) that promotions are not

beneficial (for retailers) because price promotions reduce retailer category margins.

This difference in interpretation is understandable (i.e., explainable) given that the

15

Page 23: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

margins in their equations, consistent with the literature, are not the marginal return to

shareholders. Consistent with the theory of expandable purchasing, I argue that price

promotions of a product will accelerate inventory turnover. The inventory turnover

acceleration, in turn, results in less internal capital being used over the time of the

payable schedule. Thus,

H7: Price promotions of products, in a majority of cases, will have positive effects on promoted products’ GMROSI.

As a proposed metric, GMROSI is consistent with dynamic competition

theory, as exemplified by the resource-advantage theory of competition (hereafter, R-

A theory). As indicated by Hunt (2000, p.123), “The “superior” in superior financial

performance equates with both more than and better than” in R-A theory. By

mathematical rule, increases in GMROSI often involve financial performance that is

better than and may involve financial performance that is more than GMROII. That

is, GMROSI accounts for the funding origination in the equation, and, by so doing,

permits a more accurate picture of the return to the shareholder. In GMROII, all

funding is always from internal capital. Because this is not truly the case, GMROII

presents a negatively biased view of the return to the shareholder’s inventory

investment. Thus, GMROSI will often present a “better than” scenario that more

accurately reflects the business practice of leveraging funding across different sources

(i.e., changes in the denominator). It should be noted though that because of this

combination (of activity and investment measures), the interpretation of changes in

GMROSI (vs. GMROII) is more complex and requires careful attention.

16

Page 24: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

GMROSI may involve financial performance that is “more than” because there

is no guarantee that the top line growth will occur (i.e., changes in the numerator).

Because of this, GMROSI as a measurement of organizational performance is

consistent with the relative resource costs and relative resource-produced value axis of

the competitive position matrix of comparative advantage (Hunt and Morgan 1997).

Thus, in adopting GMROSI as a dependent variable, the research also adopts a

dynamic competition perspective that can account for it.

Competitive Intensity

In regards to a dynamic competition perspective, this research will advance

understanding of the potential moderating role of competition on the relationship

between retailer promotions and retailer performance. Ailawadi et al (2006) use two

indicator (i.e., dummy) variables to represent the presence of either same-format

retailer competitors or alternative format retailer competitors. They find the two

effects to be different. This study would replicate their test in a new setting—an

every-day-low-price, general-merchandise retailer. More importantly, and drawing

upon R-A theory (e.g., Hunt and Morgan 1996; Hunt 2000), the proposed research

would investigate competitive intensity at a more detailed level, accounting for the

potential varying impact of approximately 30 different competitors that compete at

different levels (e.g., category, total-store) with the retailer.

Also, product competition occurs within a store (e.g., choosing Regular Oreos

vs. Chips Ahoy vs. Reduced Fat Oreos vs. a Snickers Candy Bar). Thus, research

should take into account the level of heterogeneity of the merchandise assortment in

17

Page 25: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

the store in which the item is price promoted. As the competitive intensity increases

either within or across stores, consumer price sensitivity increases. With increased

acuteness, consumers are better able to process price promotion information, thereby

making more informed decisions. In turn, the enhanced decision-making ability

permits consumers to expand purchasing and consumption, leading to improved

quality of life. Following the preceding discussion, I hypothesize:

H8a to H8k: Increases in merchandise heterogeneity will strengthen the relationships in hypotheses 1-7.

H9a to H9k: Increases in the number of total competitors will strengthen the

relationships in hypotheses 1-7. H10a to H10k: Increases in the number of primary competitors will strengthen

the relationships in hypotheses 1-7. H11a to H11k: Increases in the number of secondary competitors will strengthen

the relationships put forth in hypotheses 1-7. Lodish (2007, p.24) proposes that “for practitioners, the geographic differences

in market response are much more important than share differences because they are

directly actionable and can affect profitability.” The geographic differences in market

response are due, in large part, to differences in customers. According to Ailawadi et

al. (2006), the demographic customer differences include education levels, income

levels, and ethnic diversity levels. In their analysis, they find that education has a

negative effect on price elasticity, high income has a negative effect on price elasticity,

and minority ethnicity (i.e., Hispanic or Black) has a negative effect on price elasticity.

Yet, they provide no rationale for why any of these effects might occur. Recall that in

a theory of expandable consumption, there are multiple motives, one of which is

18

Page 26: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

identity. One reason why Ailawadi et al. (2006) find these effects might be that higher

income or more educated customers do not wish to be identified with advertising or

price discounting). Controlling for the effects of income and education, one potential

reason why customers of particular ethnicities might have been found to be less price-

sensitive in prior studies might be that they have less societal access to act on the price

promotions. That is, they have less capacity to consume that might be due, in part, to

their geography. For instance, they might live in areas with higher public

transportation usage (where additional product purchasing is more difficult to

transport to their residences) or higher housing density (where, as a result, house or

apartment sizes are smaller, resulting in less capacity to stockpile merchandise).

Regardless of whether capacity to consume does, perhaps, explain this ethnic, or

cultural effect, this research incorporates measurement of customer ethnicity,

consistent with prior research findings.

In summary, while the setting of this research is an every-day-low price

(EDLP), general-merchandise retailer, and not a hi-low price, drug-store retailer, there

is no reason based on the prior discussion as to why the relational signs between the

study of Ailwadi et al. (2006) and this study should be different. Rather, for a general-

merchandise retailer, these relationships should be stronger, if any differences do

exist, given that the limited, prior research finds shoppers in EDLP chains to have

higher regular or long run price sensitivities (e.g., Shankar and Krishnamurthi 1996).

Thus,

H12ato H12k: Increases in the level of customer education will have a negative effect on the relationships in hypotheses 1-7.

19

Page 27: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

H13a to H13k: Increases in the amount of customer income will have a negative

effect on the relationships in hypotheses 1-7. H14a to H14k: Increases in the percentages of African American customers will

have a negative effect on the relationships in hypotheses 1-7. H15a to H15k: Increases in the percentages of Hispanic customers will have a

negative effect on the relationships in hypotheses 1-7.

Another potential moderator of the promotion-performance relationships stated

in this research is the product’s purchase frequency. For instance, razorblades are

purchased more frequently than toasters. In either situation, this often corresponds

with, but yet is distinct from, the product durability. It is possible for durable products

(e.g., toasters) to become commodities. It is proposed here that the perceived

consumability of these products is affected by price promotions. No evidence is found

in the literature to propose that the preceding hypotheses will be different (e.g.,

positive effect, negative effect) for more frequent (i.e., health and beauty aid) products

vs. less frequent (i.e., small appliance) products. However, will the effect size be

similar, or will the effect size of one type of product be larger? Thus,

RQ2: Is the effect of price promotions in H1 through H7 greater for more frequently purchases products or for less frequently purchased products?

Last, this research project investigates when (i.e., under what competitive

conditions) customers are “cherry picking” shopping due to the price promotion versus

“self-indulging” or “hedonic” shopping due to the price promotion. That is, under

what conditions is it more likely that consumers will respond to the price promotion

by purchasing more reduced price products—reducing profitability growth versus

20

Page 28: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

sales growth (i.e., a reduction in profitability margin: sales minus profitability, divided

by sales)? Alternatively, under what conditions is it more likely that consumers will

respond the price promotions by purchasing more regularly priced products—

increasing profitability growth versus sales growth (i.e., an increase in profitability

margin)? Both situations are consistent with the theory of expandable purchasing and

consumption. The conditions investigated in this research include changing

competitive intensity and customer demographics. However, there is no research

published (based on an in-depth review of the literature) to indicate when one effect

over the other effect will be found in the data analysis of this research study. Thus, the

investigation takes a research question form:

RQ3a: When does the moderating effect of H8a to H14k decrease profitability margin for either the product or market basket?

RQ3b: When does the moderating effect of H8a to H14k increase profitability

margin for either the product or market basket?

Overview of the Research Design

The sample for this research consists of weekly, point-of-sale, stock-keeping

unit (hereafter, SKU), scanner data for manufacturer-branded products collected in all

of the approximately 450 stores of an every-day-low-price, general-merchandise

retailer operating in the northeastern and midwestern United States. Several filters

were applied to available product SKUs (that were price-promoted by the retailer) to

arrive at the final sample that is consistent with the stated research purpose to

investigate the impact of retailer price promotions on retailer performance.

21

Page 29: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

First, and central to the stated research purpose, product samples are limited to

particular products where pure price promotions (i.e., no additional promotional

activities for the investigated products) were taken at some point during the period

June 2004 to July 2006. By using only pure price promotions, the first instance of

general customer price-reduction awareness is in-store when the customer views the

shelf label (which states the prior and current price, both in dollars). This setting

contains the advantaging of preventing a needed simultaneous investigation of

customer price knowledge and the source of customer price comparison (e.g., see

Vanhuele and Dreze 2002).

Second, following the finding of Ailawadi et al (2006) that the greatest

variance (in their study) is in the general merchandise and health and beauty

merchandise categories, the product SKUs in this study are chosen from these same

merchandise categories. In contrast to both the drug store setting of Ailawadi et al

(2006) and the grocery setting of other related research (e.g., Janakiraman, Meyer,

Morales 2006; Mulhern and Leone 1991; Walters 1991; Van Heerde, Leeflang, and

Wittink 2004), however, this research uses samples drawn from an every-day-low

price, general-merchandise retailer in the United States, which permits investigation of

the generalization of their findings (to a general-merchandise store context), in

addition to the new contributions on total shopping cart consumption, shareholder

return, and competitive intensity.

Third, products are selected for which the product market baskets do not

include any of the other products being investigated in the study (i.e., unique market

22

Page 30: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

baskets). Unique market baskets are operationalized here as market baskets where the

overlap (in investigated products) is less than 1 in a 100 common occurrences. It

should be noted that while the filters decrease the level of “noise” in the sample data,

there is always the possibility that some other products may be price-promoted that

customers bought as part of some of the market baskets because of the natural field

experimental setting of this research. This does not imply, however, that the findings

of this research are overstated. Rather, it suggests that any corroboration of hypotheses

in the data analysis is understated (because the relationships are significant even in the

presence of the potential marketing mix activities on other products in the market

baskets).

In all, four product SKU samples, each of N = approximately 450 (same)

stores, are found that match the pure promotion, timeframe, merchandise category, and

unique market basket filters. Two of the product SKU samples are consumable

products from the health and beauty aid category. The other two of the product SKU

samples are more durable products from the small appliances category.

This research uses the retail store’s physical size (e.g., 122,000 square feet) as

a proxy for merchandise assortment, or the intra-organization competitive intensity,

because the number of product combinations at the store level is astronomical, making

the use of all product combinations beyond available degrees of freedom for statistical

testing. This proxy has been used in prior literature (e.g., Boatwright, Dhar, and Rossi

2004), and has been found to work because the larger the store, the larger the number

of merchandise categories and the deeper the penetration within categories.

23

Page 31: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Competitor stores of 30 potential, major competitors are geographically

identified for each store in the samples, first by census data, and then confirmed by the

store manager. This store-level competitor information is documented and retrieved

from the retailer’s scanner database. Some of the competitors compete in particular

categories, while others compete across many categories. Some of the competitors

carry higher priced products, while others carry lower priced products. This research

will account for these complexities using R-A theory as a foundation. Adopting this

foundation, the research will distinguish between category and store competitors in

modeling the complexities by (1) using indicator (i.e., dummy) variables for each

competitor and (2) identifying whether a given competitor is a total store competitor or

a category competitor. The increased sensitivity of these tests will advance

understanding of the moderating effects of competitive intensity on the relationship

between retailer promotional activities and retailer performance.

The price promotion’s effect on performance is measured against two baselines

for each product SKU sample. The first baseline consists of a moving average baseline

(e.g., Abraham and Lodish 1993). Because sufficient prior data is available, a second

baseline, consisting of prior-year, same-period data, is also used. Paired samples T-

test procedures are used to compare the difference between each baseline, in turn, and

the customer purchase behavior (i.e., product and market basket performance) after the

price promotion occurs.

In regards to comparing changes in GMROII and GMROSI, first, a simulated

response surface is constructed to provide a theoretical range of GMROII and

24

Page 32: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

GMROSI values that retailers, in general, might expect to achieve under different

variable (i.e., gross margin, inventory turnover, contracted terms of payment, and

internal rate of return) combinations. See Myers (1971) and Myers and Montgomery

(1995) for overviews of response surface methodology. Ranges for the inventory

turnover, gross margin, account payable (terms of payment), and internal rate of return

variables are developed based on analyses of several retailers’ annual reports. The

response surface permits investigation of the cost of capital while avoiding entry into

the debate over its calculation (see Chen, Dhaliwal, and Xie 2006). Second, GMROII

and GMROSI metrics are calculated for the baseline and post-promotion periods of

the four natural field experiments (each of N = approximately 450 stores). Paired

sample T-tests are used to compare pre- and post-promotion GMROII and GMROSI.

Third, these results are discussed in terms of their location on the generated response

surface.

In regards to the hypotheses related to competitive intensity, the market basket

performance (sales, profits, inventory turnover) and product GMROSI data are entered

as dependent variables in a multivariate general linear model, with competitive

intensity scenarios as fixed factors and customer demographic variables as covariates.

See Figure 1. Multiple comparisons tests (i.e., Bonferroni and Tukey honestly

significant difference) are run to investigate all pairwise comparisons between

competitive intensity scenarios, permitting corroboration of the auxiliary hypotheses

on competitive intensity.

25

Page 33: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

CHAPTER 2

A REVIEW OF THE LITERATURE

Overview

This chapter presents a review of the extant literature to the current research.

The chapter is structured as follows. Section 1, entitled “Price Promotions and

Definition,” identifies what type of promotional activity (and thus related literature),

either is or is not under inquiry in this research study. Section 2, entitled “Price

Promotions and Expandable Purchasing,” provides a general synopsis of the literature

concerning the effects of price promotions on the promoted product and its associated

market basket. Section 3, entitled “Price Promotions and Return Calculation,”

discusses the literature on return on investment metrics, as applied to retail product-

level analysis. Section 4, entitled “Moderators of the Promotion-Performance

Relationships” reviews the literature on competitive intensity, including intra-store

competition, inter-store competition, customer demographics, and product durability.

Hypotheses and research questions are presented at the end of each section. A

summary of all hypotheses is presented at the end of the chapter.

Price Promotions and Definition

Promotions can take several forms, from pure price promotions to sales

promotions (i.e., advertising). The purpose of this research is to increase our

understanding of the impact of pure, retail price promotions on retailer financial

performance. In this dissertation, a pure, retail price promotion (hereafter, price

26

Page 34: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

promotion) is defined as the temporary discounting of a product at retail where the

only indication of the discount to the customer is the stated change on the shelf label

adjoining the product (Blattburg, Briesch, and Fox 1995).

Thus, “pure” (in pure, retail price promotion) refers to the absence of any

additional displays, features, coupons, advertisements, or other intentional activities by

the retailer that could confound the effects of price promotions (see Van Heede,

Leeflang, and Wittink, 2004, for additional discussion on “pure” promotions).

“Retail” (in pure, retail price promotion) refers to price promotions taken by retailers

to consumers, not manufacturers to retailers, on products. Consequently, the analysis

adopts a retailer view of price promotions. A retailer orientation is, according to

Mulhern and Leone (1991, p. 64), “vastly different from the brand manufacturer

orientation that dominates the marketing literature.” They (1991, p. 64) state that

“typical depiction of retailers in the marketing literature as intermediaries that

distribute products from manufacturers to consumers” is an “inappropriate framework

for many marketing channels because of the relative power of manufacturers and

retailers has changed dramatically in the past few decades.” “Price” (in pure, retail

price promotion) indicates that the promotional activity is not to be confused with

sales promotion, or advertising, which includes activities such as “contests, point of

purchase displays, sampling premiums, coupons, multi-package price deals, incentive

programs, tie-in sales, and certain forms of direct mail” (Adler, 1963, p. 69).

“Promotion” (in pure, retail price promotion) refers to the dissemination of

27

Page 35: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

information from retailer to customers through the shelf label that the product has been

temporarily discounted to the retail customer to further the goals of the organization.

Thus, pure, retail price promotions are distinct from products that are retailer

advertised but not retailer price discounted (e.g., to create awareness of the product

availability at the store). They are also distinct from products that are retailer

advertised and retailer priced discounted (e.g., to increase customers visiting the

store). Pure, retail price promotions, in contrast, increase purchasing volume across

customers and time (e.g., Blattberg and Wisniewski 1987; Duncan, Hollander, and

Savitt 1983; Kuehn and Rohloff 1967; Mason and Mayer 1984; Moriarty 1985;

Woodside and Waddle 1975), and are usually the most frequent type of promotion

used by retailers (e.g., Gedenk, Neslin, and Ailawadi 2006), especially for every-day-

low-price retailers (e.g., Walton and Huey 1992).

As by indicated by Blattberg, Briesch, and Fox (1995, p. 122), “the price

promotions literature is new relative to other research in marketing, having been

developed primarily since the early 1980s.” Despite its relative newness, the subject of

price promotions is an important, and consequently, extensively researched area of

marketing thought. In the words of Dekimpe et al. (2005, p. 409), “Price promotions

are the most often used form of promotional support. As such, it should come as no

surprise that the effectiveness of price promotions has been studied extensively in the

marketing literature.” However, as noted by Assuncao and Meyer (1993, p. 517),

although the question “How should price promotions affect the purchase and

28

Page 36: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

consumption of goods…lies at the heart of much of the modern literature on sales

promotions, it is one for which we currently do not have a complete answer.”

Consistent with both the (1) the importance of this research stream and (2) the

paucity of current information about how price promotions affect customer

purchasing, the purpose of this research is to increase our understanding of the impact

of pure, retailer price promotions. The underlying research question of this study is:

What is the impact of retailer price promotions on retailer performance? Specifically,

I focus on price promotions and two types of retailer performance: (1) the retailer’s

total market basket performance (i.e., market basket sales, profitability, and turnover),

and (2) the price-promoted products’ financial return on shareholder investment. A

causal model indicating the hypotheses to be explored is presented in Figure 1.

Prior literature has found that retail price promotions have a positive effect on

product sales volume (i.e., units sold) and dollars (i.e., gross sale dollars) during the

promotional period (e.g., Ailawadi and Neslin 1998; Blattberg and Wisniewski 1987;

Moriarty 1985; Van Heerde, Gupta, and Wittink 2003; Woodside and Waddle 1975).

For listings of the many articles reporting the elasticities and cross elasticities of

product price promotions on products/brands and substitutes in same categories

(secondary demand) and longitudinal/early purchasing (primary demand), see table

summaries in Bell, Chiang, and Padmanabhan (1999, p. 510), Mulhern and Leone

(1991, p. 67), or Van Heerde, Gupta, and Wittink (2003, p. 483).

The increase in sales volume (due to the price promotion), in turn, also

typically accelerates the inventory turnover rate. However, when retail price

29

Page 37: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

promotions increase sales volume, how often (or when) is the increase sufficient to

improve profitability (i.e., initial gross margin dollars)? For instance, a price

promotion of Oreo’s could result in $50 sales growth, a $10 gross margin loss, and a

one-fold increase in inventory turnover. Or, it could result in a $100 sales growth, a

$20 gross margin growth, and a two-fold increase in inventory turnover. Given the

substantial, previous research on the effects of price promotions on sales volume and,

in contrast, the paucity of previous research on the effects of price promotions on

profitability, I hypothesize (and ask):

H1: Price promotions of a product will have a positive effect on that product’s sales (at the product-store level).

H2: Price promotions of a product will have a positive effect on that product’s

inventory turnover (at the product-store level).

RQ1: What is the effect of price promotions of a product on that product’s profitability (at the product-store level)?

Price Promotions and Expandable Purchasing

As indicated in the research question, one contribution of this research is to

provide the first documented measurement of the impact of pure price promotions of

individual products on the total purchase of the customer per shopping visit—or

market basket. The market basket is defined here as the total products acquired by a

customer during a single purchase event at the store checkout register (i.e., the total

“shopping cart” purchased). According to Hansen, Raut and Sumit (forthcoming, p.1),

“research is needed that investigates the effect of price promotions and advertising on

market basket performance.” While some research has begun to explore the potential,

30

Page 38: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

particular product, complement (e.g., inter-category performance) and product

substitute (e.g., intra-category performance) effects in market baskets (e.g.,

Janakiraman, Meyer, and Morales 2006; Shocker, Bayus, Kim 2004), no study has

looked at the impact of individual product pure price promotions on retailer total-

market-basket sales, profitability, and inventory turnover. See Table 1. The absence of

research on total-market-basket effects could be because of the difficulty of acquiring

data, which is usually available only through proprietary, store-level scanner data (see

Bucklin and Gupta 1999). Consider, for instance, the research sample design of the

following studies.

There is a growing stream of literature that finds that retailer price promotions

can not only increase the purchase of promoted product(s), but also can result in

changes in the purchases of nonpromoted, substitutable and complementary products

(e.g., Ailawadi et al. 2006; Ailawadi and Neslin 1998; Chandon and Wansink 2002,

Chintagunta and Haldar 1998; Janakiraman, Meyer, and Morales 2006; Mulhern and

Leone 1991; Walters and MacKenzie 1988; Walters 1991). This research stream is

recent, with many researchers citing MacKenzie and Walters (1998), Mulhern and

Leone (1991), and Walters (1991) as the first to investigate the potential effects of

product price promotions on other products.

Walters and MacKenzie (1988) state that in-store price promotions have no

effect on the sales “traffic” (i.e., the number of transactions), using store scanner data

from two stores of a large midwest supermarket chain on sales of cake and ready

made frosting and spaghetti and spaghetti sauce. They state “(1988, p. 52), “Unlike

31

Page 39: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

manufacturers, grocery retailers are not as interested in the effect of price promotions

on individual product performance as they are in the effect of marketing and product

mix changes on overall store performance.” Their findings should be interpreted in

view of their product sample selection criteria. First, “the product must have received

a reasonable amount of promotional attention during the data collection period. For

example, gourmet candy, private label crackers, and oven cleaner were promoted only

once or twice a year so few observations were available on those items. Consequently

such products were excluded from the analysis” (p. 56). In contrast, the sample design

of this research uses product SKUs that are at most promoted once or twice a year.

Also, they use structural equation analysis of instore price promotions to explain

variance in total store sales by week.

Mulhern and Leone (1991) analyze weekly POS (sales & profits figures) for

cake and cake frosting categories for a regional grocery store chain. They estimate the

cross category effect using cross-elasticity parameter estimations. Walters (1991)

investigates complementary relationships between cakes and cake frosting and

spaghetti and spaghetti sauce and the substitutes in each category in (8 pairs in 8

markets) 64 stores in the midwestern US, developing cross elasticities measuring total

category sales for each category (i.e., not necessarily same basket purchases). There

is, consistent with most other research in this area, no control for advertising or other

marketing mix activities.

Mulhern and Padgett (1995), survey shoppers to determine whether advertising

products (of a flyer containing 200 featured products) drive sales and profits dollars

32

Page 40: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

for two stores of a home improvement chain of a dozen stores. Retail clerks administer

the surveys to customer in two of 11 stores in a home improvement company. They

analyze 412 surveys and find that the surveyed promotion shoppers spend more

money on regular priced merchandise than on promotion merchandise. According to

their analysis, shoppers who did not visit the home improvement retailer because of

the promotions but did purchase promoted products, purchased $33.90 of merchandise

(resulting in $5.74 in profit for the retailer), on average. In contrast, shoppers who did

not purchase promoted products purchased $23.10 of merchandise (resulting in $5.85

in profit for the retailer), on average. According to their analysis, 13 of 56 (23.2%)

shoppers buying promoted products “cherry picked.” That is, they only purchased the

promoted products (and nothing else). However, the profitability in their study is

“computed by subtracting unit cost from unit price and summing unit margins across

all products purchased by each customer.”

Chintagunta and Haldar (1998) use household level data to investigate

complementary purchases of two categories using three “consumable” pairs: pasta and

sauce, liquid and powder detergents, and soup and yogurt (with both promoted), as

well as two “durable” pairs: and clothe washers and clothe dryers, and dishwashers

and air conditioners. They (p. 53) state that “we recognize that we have merely

scratched the surface in terms of investigating cross-category effects. It would be

necessary to extend our analysis to multivariate hazards (more than two

categories)…Furthermore, it would be important to examine individual brands rather

than the category as a whole.” In the same year, Ailawadi and Nelson (1998, p. 396)

33

Page 41: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

look at price, but not price promotions, using household scanner data. There is no

control for other promotional activities.

Nijs et al. (2001) use IRI household scanner data covering 560 categories in

the Netherlands to investigate the effects of different marketing mix elements,

including price promotions and advertising, all present at the same time, of two

categories on 79 other food categories using household scanner data. They (p. 10) state

that competitive structure is “measured by the number of brands in each category. This

is one of the best known measures of competitive market structure.” In contrast, by

adopting a retailer perspective, competitive structure is investigated in this dissertation

at the retailer level (as opposed to the brand level).

Bell, Chiang, Padmanabhan (1999, p. 514) use a sample of 250 IRI panelists in

three supermarkets over 78 weeks to review 13 categories of grocery and consumable

paper goods. Likewise, Van Heerde et al (2004, p. 326) use ACNielson store level

scanner data at the brand level for four product categories (i.e., tuna, tissue, shampoo,

and peanut butter) in stores in the US and two categories (i.e., shampoo and peanut

butter) in stores in Denmark to estimate price promotion effects. Across the categories,

they find an own-brand effect of .44, cross-brand effect of .35, cross period effect of

.44, and category expansion effect of .2. They (p.. 332) conclude that “From a retailer

perspective, a 100-unit increase for the promoted brand causes on average a net 33-

unit decrease for other brands in the category and a net 32-unit decrease in pre- and

post-promotion category sales (Table 4). This leaves a potentially beneficial 35-unit

increase for the retailer due to category expansion effects.”

34

Page 42: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Bell, Iyer, and Padmanabhan (2002) use store level scanner data from several

hundred products spread across eight product categories in five stores in one market.

They find that consumption can be flexible, incremental as people stockpile due to

price promotion. Shocker, Bayus, and Kim (2004, p. 32) propose that “in-use” and

“occasional” substitutes and complements are connected and that price promotions

have an effect on them. They (p. 37) propose that one future research direction is to

address “What are additional managerial implications from influences of ‘other

products’?” One managerial implication of the current research study is that market

baskets are comprised of the product, its complements, its substitutes, and ‘other

products’ (or in the words of Shocker et al. “occasional” substitutes and complements)

that are random components. Other products may, in fact, represent much of the

potential market basket increase that comes from hedonic shopping. Thus, a total

market basket orientation, versus an orientation toward category substitutes or

commonly occurring complements, provides a more accurate view of the effect of

price promotions on retailer performance.

As a consequence of this new and growing literature on price promotions and

market basket analysis, there are calls for price-promotion research to move from a

product or brand focus to a category management level of analysis (e.g., from focusing

on Oreos to focusing on cookies; see Nijs et al. 2001). “The primary implication of

category management,” according to Bucklin and Gupta (1999, p. 265), “is a shift in a

retailer’s attitude from a buyer orientation (where money is made on how the product

is bought), to a buyer and merchandiser orientation (where the focus is on category

35

Page 43: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

profitability.” Research on, and discussion of, category management typically refers to

intra-category management, or management of one category of market offerings. This

type of category management is a part of, but not equivalent to, market basket

management.

Market Basket Management

Market basket management involves both intra-category and inter-category

performance management. In regards to intra-category performance, a product price

promotion can affect, for instance, substitute products (e.g., competing brands of hot

dogs). In regards to inter-category performance, the product price promotion can

affect, for instance, complementary products (e.g., hot dogs, hot dog buns, and

condiments). While a retail manager (based on typical reward systems) is most

interested in category analysis, the retailer (as an organization) is most interested in

market basket effects that are closer to representing the total store performance.

Consequently some researchers argue for price-promotion research to shift

even further, from category management to investigating total store purchases (e.g.,

Bucklin and Gupta 1999; Shocker, Bayus, and Kim 2004). In particular, two recent

articles have provided initial evidence implying that price promotions can increase

total shopping cart (i.e., store-level, market basket) purchases per shopping visit

(Ailawadi et al. 2006; Janakiraman, Meyer, Morales 2006).

For example, the recent work by Ailawadi et al. (2006, p.526) is “the first to

estimate halo [i.e., market basket expansion] rates.” However, they use a regression

36

Page 44: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

based estimate of the quantity of items in the market basket based on a sample of store

loyalty card information because they do not have access to market basket information

for their store-level scanner data. They then estimate the market basket sales dollars

and gross profit dollar by multiplying the estimated change in units by the total store

sales and profits averages. There are several limitations to such store-level estimations,

one of which is the inability to calculate the market basket metrics for each individual

promotion (and individual promotions are, indeed, the specified unit level of analysis).

Therefore, Ailawadi et al (2006, p. 531; emphasis added) call for future researchers to

“validate our work, find new ways of obtaining more disaggregate estimates of the

halo effect, and study how and why halo effects vary across categories and retail

formats.”

In contrast, Janakiraman, Meyer, and Morales (2006, p. 363) use an

experimental design, involving the participation of 150 undergraduate students in an

computerized shopping simulation of 12 products over 35 weeks. They find positive

“spillover effects” of discounted products on the purchase of other products during the

same simulated shopping trip. Combined with the findings of Ailawadi et al. (2006),

these results indicate that price promotions affect more than the promoted product, and

thus, should be analyzed, and perhaps managed, at a more strategic level. However,

despite the importance of investigating more strategic perspectives of retailer

performance, “relatively few studies have focused on cross-category price-promotion

effects, especially at the retail store level” (Kamakura and Kang (2006, p. 159).

37

Page 45: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

The purpose of this research is to model the effects of the price promotions on

the retailer (i.e., the organization)—not the individual retail manager. Adopting this

more strategic perspective, I focus on the broader market basket perspective as

opposed to the narrower category management perspective. Thus, furthering the goal

of understanding the effect of promotional activities on retailer performance, this

study is the first study to measure, instead of estimate, the effect of individual pure

price promotions on the actual, total-market-basket performance. See Table 1. Prior

studies, in contrast, must estimate dependencies across products and categories using,

for example, elasticities (e.g., Manchanda et al. 1999; Mulhern and Leone 1991;

Walters 1991) hazard models (e.g., Chintagunta and Haldar, 1998), or multiple

equation time series models (Dekimpe and Hanssens 1995; Nijs et al. 2001; Srinivasan

et al. 2000).

Toward a Theory of Expandable Purchasing and Consumption

Synthesizing the findings of prior research on the subject of market basket

(e.g., Ailawadi et al 2006; Janakiraman, Meyer, and Morales 2006; Mulhern and

Padgett 1995), this research proposes that purchase and consumption are expandable.

One example is the “halo effect” proposed by Ailawadi et al (2006). Another example

is the “spillover effect” proposed by Janakiraman, Meyer, and Morales 2006. Other

examples are found in the works of Chandon, Wansink, Laurent (2000), who identify

six major multiple consumer benefits from sales promotions (that could also apply to

price promotions): opportunities for value expression, entertainment, exploration,

38

Page 46: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

savings, higher product quality, improved shopping convenience. Also, McCracken

(1999) presents several other potential reasons for consumption, including identity,

emotional fulfillment, and a consistent product constellation. This dissertation

proposes that (what I refer to as) expandable purchasing can occur as a result of any of

these, or perhaps others, individually or in combination.

The word “expandable” indicates that, contrary to the commonly-accepted

“market pie” metaphor in marketing (e.g., Carson et al. 1999; Chakravarti and

Janiszewski 2004; Day and Montgomery 1999; Jap 1999; Nason 2006; Rokkan,

Heide, and Wathne 2003; Weitz and Bradford 1999), consumption is not like a static

“market pie” to be divided (i.e., maintained). This usage of the pie metaphor is an

adaptation of the neoclassical economic pie metaphor. Speaking on the original usage

of the metaphor in macroeconomics, Swinnerton (1997, p. 75) states, “To develop a

better intuition for the definitions of efficiency and their equivalence, it is useful to

think of a pie as a metaphor for the output of the economy.” In the pie metaphor in

marketing research, however, researchers have equated pie size with “primary

demand” (instead of labor allocation) and pie distribution with “market share” (instead

of wealth distribution). The pie metaphor, as currently applied to purchasing-related

research, implies that all consumers go to a store with a budget and they spend the

budget. This adaptation of the metaphor has resulted in interpretations of price

promotions as zero-sum games of product switching (e.g., Dodson, Tybout, and

Sternthal 1978).

39

Page 47: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

The growing literature converging on a theory of expandable purchasing does

not support this metaphor or its resulting interpretation of price promotion research.

Instead, this growing stream of research is more consistent with a balloon metaphor,

where purchasing can be grown similar to the inflating of a balloon (Hansen and

McGinty, 2007). Rather than assuming that all consumers go to a store with a budget

and spend the budget limit, this approach assumes that consumer may visit with a

budget in mind, but retain the flexibility to adjust spending. For instance, a consumer

may plan on spending 100 dollars as he has done in the past, but decides to spend

more because of the effect of the pure price promotion. Similar to the balloon analogy,

a shallow discount will not result in much of an effect on overall purchase behavior.

However, the effect grows nonlinearly with the depth of the price promotion. That is,

as the size of the price promotion for a product increases, there is expanded purchasing

of other products. These products may be complements or may be completely

unrelated. Indeed, according to Walters and MacKenzie 1988, p. 54), “Retailers and

marketing researchers generally believe one of the primary benefits of price

promotions is that they stimulate sales not only for the lower price, lower margin

promoted items but also for higher margin goods that are not being promoted.” While

the effect is proposed to be nonlinear, hypotheses here are only directional due to

normal data limitations. That is, longitudinal data on particular products over several

promotional prices is needed to compute the nonlinear elasticities. Such data is rare,

not present in the literature, and not available for this research study.

40

Page 48: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

The motivation for expanded purchasing may be economic utility

maximization, hedonic enjoyment, a reward to the retailer, or a combination of these

and other consumer motivations. I propose that there is a point (though it might be

difficult to locate precisely) where a very deep price promotion might result in

negative reaction by consumers who begin to doubt the quality of the product given

the large price reduction (similar to the balloon popping). See Kirmani and Rao (2000)

for discussion on signaling unobservable product quality. Most retailers, however, do

not engage in this kind of price promotion activity, with the exception of occasional

loss leaders (e.g., milk, bread, crayons, beer). Rather, the relevant range of the type of

retail price promotions being investigated (i.e., context) is expected to be in the “front

half,” or positive slope area, of the proposed balloon effect. Based on the preceding

discussion, I hypothesize:

H3: Price promotions of a product will have positive effects on market basket sales.

H4: Price promotions of a product will have positive effects on market basket

item count (i.e., number of products in the market basket).

H5: Price promotions of a product will have positive effects on market basket profitability.

Price Promotions and Return Calculation

Another contribution of this research is that it investigates the effect of price

promotions on shareholder investment. There have been calls (e.g., Lehmann 2004;

Rust et al. 2004) for marketing research to build bridges between marketing activities

and financial (e.g., shareholder) outcomes. According to Srivastava, Shervani, and

41

Page 49: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Fahey (1998, p. 2), marketers must “understand the financial consequences of

marketing decisions, expanding the external stakeholders of marketing to include

explicitly the shareholders and potential shareholders of the firm.” In their words,

moving from the tradition assumption of sales, ROS, and equity as operational

measures toward the emerging assumptions of net present value of cash flow and

shareholder value.

It is argued here that price promotions of a product can make other resources

(i.e., internal capital) of an organization more productive. While Kumar, Madan, and

Srinivasan (2004, p. 933) state that “Price promotions commonly increase

manufacturer revenue and depress retailer revenue in the short term but have no

persistent effect. Promotions are tactical, not strategic, and they need to be managed

that way.” While the effect of a promotion is, according to the literature, often short-

term, the usage of promotions over time carries strategic implications. As indicated

prior, by adopting a retailer—rather than a retail manager—perspective on the

effectiveness of price promotions, this research advances the price promotions

literature toward strategy formulation, answering the calls of researchers such as Day

(1992) and Webster (1981, 1992).

Further, recent accounting research recognizes and calls for research that

addresses the difference between ROI and the real economic profitability of an

organization (see Rajan, Reichelstein, and Soliman, forthcoming). This research will

be the first study, based on a review of the literature and discussion with several

42

Page 50: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

thought leaders in the area, to quantify the impact of retailer promotional activity on

the retailer’s shareholders by introducing and using a new metric.

Early works on valuation often used return on investment (ROI).

Unfortunately, the “investment” in ROI varies by research study. The formula

variation (in studies purporting to use ROI) has resulted in several noted potential

problems in interpretation. Most often ROI appears to become return on sales—

which should be identified as ROS, not ROI. ROS is calculated as:

(1) ROS = Net Income (Before Interest & Tax) ÷ Sales Dollars

Practitioners are usually more interested in measuring performance within their

organization (e.g., particular product SKU price promotions) versus measuring

performance across organizations. Attempting to apply ROI (or ROS) to product (or

even category) merchandise decisions is, as indicated by Sweeney (1973, p. 61),

“fraught with cost measurement and allocation problems.” Further, price promotions,

by mathematical rule, decrease the unit margin in order to stimulate sales growth. In

the context of ROS, the growth in the numerator (i.e., net income) is less than the

growth in the denominator. Take, for example, a box of Tide detergent that is regularly

prices at $6, which generates $2 in net income and a ROS of 2/6=33%. The Tide is

price promoted at $5 per box, which decreases net income to $1 and results in a ROS

of 1/5=20%. While there may be a ten-fold increase in sales volume, increasing both

the sales dollars and income dollars, the ROS ratio decreases. Thus, neither ROI nor

ROS is well equipped to measure product-level, operational performance. However,

both the allocation and dollars versus percentage problems can be overcome when

43

Page 51: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

“the performance measure is used exclusively for planning and controlling

merchandising inventory investment” (Sweeney 1973, p. 61).

Consequently, a few academic researchers and many practitioners have

adopted a more specific measure of return, the gross margin return on inventory

investment (hereafter, GMROII) to plan and control (e.g., Ahern and Romano 1979;

Dunne and Lusch 2005; Kravitz 1977; Leeds 1976; Sweeney 1973; Tolle 1976;

Warrington 1982). GMROII is calculated as:

(2) GMROII ( % ) = Gross Margin (%) ÷ [1 - Gross Margin (%)] x Inventory Turnover

One advantage of GMROII is that it does not require any approximations from

aggregated activities such as labor costs. Thus, it can be applied at the product,

category, or overall retailer level of analysis. Another advantage of GMROII (over

ROS) is that GMROII accounts for inventory turnover. Also, while other return

metrics (e.g., ROI, ROS) do involve a time component (i.e., it is the return on

investment or sales for a given period of time), GMROII emphasizes the importance

of time. That is, it shows how time can provide a strategic advantage/disadvantage as

an organization manages its inventory (i.e., cost of goods sold) better/worse than its

competitors. Despite these advantages, there is a dearth of academic literature that

uses GMROII. Perhaps one reason why few academic researchers use GMROII (in

contrast to many practitioners) is access to data on both profitability and inventory

turnover rates. Thus, many studies have incorporated the effects of price promotions

on consumer inventory (e.g.., Assuncao and Meyer 1993; Bucklin and Lattin 1991;

44

Page 52: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Chintagunta 1993; Guadagni and Little 1987; Gupta 1988, 1991; Nelson and Stone

1996), but none of these studies analyzes, incorporates, or even mentions the potential

effects of price promotions on retailer inventory. One result is that the effect of the

price promotions of a product on GMROII has not been established in the literature.

This effect could be either positive or negative, depending on how the gross margin

percent and the inventory turnover are conjointly affected by the price promotions.

For example, a price promotion of widgets might result in both a ten percent

decrease in gross margin percent (i.e., from 50 percent to 40 percent) and a three fold

increase in inventory turnover (i.e., from two to six turns annually). Or, the same price

promotion of widgets might result in both a five percent decrease in gross margin

percent (i.e., from 50 percent to 45 percent) and a one fold increase in inventory

turnover (i.e., from two to three turns annually). The first scenario (smaller gross

margin, larger turnover) results in a GMROII of 400 percent return, whereas the

second scenario (larger gross margin, smaller turnover) results in a GMROII of 245

percent return. . In this example, the price promotion that resulted in less of a margin

decrease did not result in a higher GMROII (because of the relative inventory turnover

impact). Thus, in hypothesizing the impact of price promotions on GMROII, a

researcher is implicitly hypothesizing about the relative change in gross margin versus

inventory turnover rate. The neoclassical economic “pie” metaphor (presented earlier

in this chapter) maintains that the inventory rate is maintained over time or across

products (i.e., substitution effects). In contrast, a theory of expandable purchasing

suggests that the inventory turnover rate would be accelerated due to the product’s

45

Page 53: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

price promotion. Based on the preceding discussion, and consistent with the growing

literature supporting the phenomenon that I refer to as expandable purchasing, I

hypothesize:

H6: Price promotions of products, in a majority of cases, will have positive effects on promoted products’ GMROII.

As a derivation of ROI, though, GMROII does not consider the financing of

the cost component, and thus retains the inability to document the impact of price-

promotions on an organization’s owners. The cost component consists of both

financing and operating responsibilities, and operational metrics such as ROI and

GMROII follow managerial accounting guidelines that have traditionally held that:

Managers have both financing and operating responsibilities. Financing responsibilities relate to how one obtains the funds needed to provide for the assets in an organization. Operating responsibilities related to how one uses the assets once they have been obtained. Both are vital to a well-managed firm. However, care must be taken not to confuse or mix the two when assessing the performance of a manager. (Garrison and Noreen 2003, p.773)

The foundational logic of this research agrees with the premises of these

quoted guidelines. That is, there are financing and operating responsibilities (i.e.,

premise 1) and both responsibilities are vital to the organization (i.e., premise 2).

However, Anderson (1979, p. 325?) states that too often marketing tends to “fail to

recognize the impact of marketing decision on such variables as inventory levels,

working capital needs, financing costs, debt-to-equity-rations, and stock prices.”

According to Srivastava, et al. (1998, p. 11), “the impact of marketing activities on the

fixed and working capital requirements of the firm, though it has received some

attention lately, general is not well understood.” It is proposed here that the inventory

46

Page 54: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

turnover rate does have an effect on the accounts payables. That is, the inventory

turnover rate can change how one obtains the funds (e.g., what is funded internally, or

funded externally, or, even, never requires funding at all).

In contrast to ROI, return on equity (ROE) considers the financial leveraging

of funding (see Block and Hirt 2000, p. 56):

(3) Return on equity = Return on assets (investment) ÷ (1-Debt/Assets)

However, the ROE metric requires valuation assumptions about dividing the costs of

labor, space, and other activities to arrive at net income that are not easily allocated to

product level analysis. Indeed, as indicated by Srivastava, Shervani, and Fahey (1998,

p. 8), “the valuation of assets is controversial.” At the same time, the cost of financing

versus investing is being debated (see Baker, Ruback, and Wurgler 2007).

Thus, an alternative metric in use is proposed here that is similar to the Du

Pont model of ROE (see Garrison and Noreen 2003, p. 70), and can be used as a

framework for discussing promotional impact on the organization. This metric

replaces net income with gross margin (i.e., the purchased cost of goods cost, prior to

allocating labor, building, and other costs) and, in accounting for the change in the

financing plan, also incorporates the internal rate of return on the unused assets. Thus,

the metric becomes the gross margin return on shareholder investment (hereafter,

GMROSI) of the product. As a planning and controlling metric, GMROSI could be

used as a working template for managers interested in attempting to strategize on the

shareholder value-planning approach (e.g., Day and Fahey 1988; Kim, Mahajan, and

Srivastava, 1995; Rappaport, 1986).

47

Page 55: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

In relation to GMROII, the GMROSI metric replaces the “inventory”

investment with “shareholder” investment by accounting for the change in the cost due

to changes in account payables that are, in turn, due to changes in the contracted terms

of payment (e.g., 2-10-net 60) between the retailer and the manufacturer of the

product, and the actual (i.e., “positive,” not “normative”) internal rate of return in the

organization. By doing so, this planning and controlling metric accounts for the

relative return on each shareholder dollar invested in the organization. See Appendix 1

for an example. Also, while GMROSI provides an operational accounting of

shareholder inventory investment productivity, it should not be confused with

shareholder market return. The relevant formulas for GMROSI computation are:

(4) GMROSI ( % ) = Adjusted Gross Margin (%) ÷ [1 – Adjusted Gross Margin (%)] x Inventory Turnover, where

(5) Adjusted Gross Margin = (Sales $ - Adjusted Cost $) ÷ Sales $, where

(6) Adjusted Cost $ = (Cost Schedule * Cost%) - ( IRR * Cost Schedule

* [1-Cost%] )

Taking a shareholder perspective, it is proposed that many promotions that

decrease the percent profitability of the product for the retailer may, in fact, increase

the percent profitability of the product for the shareholder. The argument here runs

contrary to the findings of Srinivasan et al. (2004, p. 617) that promotions are not

beneficial (for retailers) because price promotions reduce retailer category margins.

This difference in interpretation is understandable (i.e., explainable) given that the

margins in their equations, consistent with the literature, are not the marginal return to

shareholders. Consistent with the theory of expandable purchasing, I argue that price

48

Page 56: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

promotions of a product will accelerate inventory turnover. The inventory turnover

acceleration, in turn, results in less internal capital being used over the time of the

payable schedule. Thus,

H7: Price promotions of products, in a majority of cases, will have positive effects on promoted products’ GMROSI.

As a proposed metric, GMROSI is consistent with the resource-advantage

theory of competition (hereafter, R-A theory). As indicated by Hunt (2000, p.123),

“The “superior” in superior financial performance equates with both more than and

better than” in R-A theory. By mathematical rule, increases in GMROSI often

involve financial performance that is better than and may involve financial

performance that is more than GMROII. That is, GMROSI accounts for the funding

origination from in the equation, and, by so doing, permits a more accurate picture of

the return to the shareholder. In GMROII, all funding is always from internal capital.

Because this is not truly the case, GMROII presents a negatively biased view of the

return to the shareholder’s investment. Thus, GMROSI will often present a “better

than” scenario that more accurately reflects the business practice of leveraging funding

across different sources (i.e., changes in the denominator). It should be noted though

that because of this combination (of activity and investment measures), the

interpretation of changes in GMROSI (vs. GMROII) is more complex and requires

careful attention.

GMROSI may involve financial performance that is “more than” because there

is no guarantee that the top line growth will occur (i.e., changes in the numerator).

Because of this, GMROSI as a measurement of organizational performance is

49

Page 57: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

consistent with the relative resource costs and relative resource-produced value axis of

the competitive position matrix of comparative advantage (Hunt and Morgan 1997).

Thus, in adopting GMROSI as a dependent variable, the research also adopts a

dynamic competition perspective that can account for it.

Moderators of the Promotion-Performance Relationships

In regards to a dynamic competition perspective, this research will advance

understanding of the potential moderating role of competition on the relationship

between retail price promotions and retailer performance. According to Walters (1991,

p. 18, emphasis added), “Though little is known about the breadth and depth of brand

substitution and complementarity across product categories, even less is known about

promotional effects across stores.” Related, Boatwright, Dhar, Rossi (2004), propose

that retail competition, account retail strategy (e.g., EDLP pricing strategy, store size,

chain size), and customer demographics are important factors in investigating

promotional response. They use account-level (i.e., store-market combinations, as

opposed to individual stores in markets) data representing 35 Nielson SCANTRACK

markets (because, in their words (2004, p. 173), “it is the policy of both IRI and

Nielson not to release store level information”). Despite the absence of store-level

data, they (p. 182) still that “consumer demographics are very important in explaining

variation in promotion response.” They also find that retail competition accounts, in

contrast, for only three to four percent of the variation in price sensitivity in their data

(Boatwright, Dhar, and Rossi (2004, p. 186). Recall, however, that they only had

50

Page 58: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

access to account-level data. Taking a different approach, Ailawadi et al. (2006) use

two indicator (i.e., dummy) variables to represent the presence of either same-format

retailer competitors or alternative format retailer competitors. They find the two

effects to be different. This study will replicate their test in a new setting—an every-

day-low-price, general-merchandise retailer.

More importantly, and drawing upon R-A theory (e.g., Hunt and Morgan 1996;

Hunt 2000), the proposed research would investigate competitive intensity at a more

detailed level, accounting for the potential varying impact of approximately 30

different competitors that compete at different levels (e.g., category, total-store) with

the retailer. I propose that the effect of retail competition will be different here than in

prior work because (1) store-level scanner data is used (vs. account-level scanner

data), and (2) the model does aggregate all of the competitors together (vs. aggregating

competitors to either a primary and secondary indicator category, limiting any

investigation into the intensity, or count, of the number of competitors). Competitors

are also identified as primary competitors or secondary competitors in the model. A

primary competitor competes with the retailer across most product categories, or the

total store (e.g., Sam’s Club competing with Costco or J.C.Pennys competing with

Dillards). A secondary competitor competes with the retailer mostly within the

category of the product under investigation (e.g., Barnes and Nobles competing with

Target in books).

Also, competition occurs within a store (e.g., choosing Regular Oreos vs.

Chips Ahoy vs. Reduced Fat Oreos vs. a Snickers Candy Bar). Thus, research should

51

Page 59: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

take into account the level of heterogeneity of the merchandise assortment in the store

in which the item is price promoted. As the competitive intensity increases either

within or across stores, consumer price sensitivity increases. With increased acuteness,

consumers are better able to process price promotion information, thereby making

more informed decisions. In turn, the enhanced decision-making ability permits

consumers to expand purchasing and consumption. Following the preceding

discussion, I hypothesize:

H8a to H8k: Increases in merchandise heterogeneity will strengthen the relationships in hypotheses 1-7.

H9a to H9k: Increases in the number of total competitors will strengthen the

relationships in hypotheses 1-7. H10a to H10k: Increases in the number of primary competitors will strengthen

the relationships in hypotheses 1-7. H11a to H11k: Increases in the number of secondary competitors will strengthen

the relationships put forth in hypotheses 1-7. Lodish (2007, p.24) proposes that “for practitioners, the geographic differences

in market response are much more important than share differences because they are

directly actionable and can affect profitability.” The geographic differences in market

response are due, in large part, to differences in customers. According to Ailawadi et

al. (2006), the demographic differences include education levels, income levels, and

ethnic diversity levels. In their analysis, they find that education has a negative effect

on price elasticity, high income has a negative effect on price elasticity, and minority

ethnicity (i.e., Hispanic or Black) has a negative effect on price elasticity. Yet, they

provide no rationale for why any of these effects might occur. Recall that in a theory

52

Page 60: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

of expandable consumption, there are multiple motives, one of which is identity. One

reason why Ailawadi et al. (2006) find these effects might be that higher income or

more educated customers do not wish to be identified with advertising or price

discounting. Through this behavior, and, related, by avoiding shopping at price

discount stores, they might feel that they have achieved some degree of success or

financial independence in life (e.g., they don’t have to shop at, for example, a Wal-

Mart). Alternatively, Boatright, Dhar, and Rossi (2004), finding similar effects from

proxies of household wealth proxies, attribute it to differences in customer time

valuation (e.g., customers that earn more place a greater dollar figure on time, and thus

are less price sensitive. The conditions under which one rationale (over the other) is

present cannot be measured through the store-level scanner data in this study or prior

work, and is left for future research to investigate.

In regard to customer ethnicity, while Ailawadi et al. (2006) describe an effect

without discussion, Boatwright Dhar, and Rossi (2004, p. 188) state, “We have not

elected to include ethnicity in our results reported above as there is no real theory that

can account for effects of ethnicity once wealth and household composition is

controlled for. In specifications not reported here, we found that Hispanic and black

consumers are less price responsive.” Controlling for the effects of income and

education, one potential reason why customers of particular ethnicities might have

been found to be less price-sensitive in prior studies might be that they have less

societal access to act on the price promotions. That is, they have less capacity to

consume that might be due, in part, to their geography—which is related to, but

53

Page 61: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

distinct from the other variables (i.e., income, education, home value, and household

size).

For instance, they might live in areas with higher public transportation usage

(where additional product purchasing is more difficult to transport to their residences)

or higher housing density (where, as a result, house or apartment sizes are smaller,

resulting in less capacity to stockpile merchandise). Regardless of whether capacity to

consume does, perhaps, explain this ethnic, or cultural effect, this research

incorporates measurement of customer ethnicity, consistent with prior research

findings.

In summary, while the setting of this research is an every-day-low-price,

general-merchandise retailer (vs. a hi-low price, drug-store retailer), there is no reason

based on the prior discussion as to why the relational signs between their study and

this study should be different. Rather, for a general-merchandise retailer, these

relationships should be stronger, if any differences do exist, given that the limited,

prior research finds shoppers in EDLP chains to have higher regular or long run price

sensitivities (e.g., Shankur and Krishnamurthi 1996). Thus,

H12ato H12k: Increases in the level of customer education will weaken the relationships in hypotheses 1-7.

H13a to H13k: Increases in the amount of customer income will weaken the

relationships in hypotheses 1-7. H14a to H14k: Increases in the percentages of African American customers will

weaken the relationships in hypotheses 1-7. H15a to H15k: Increases in the percentages of Hispanic customers will have a

weaken the relationships in hypotheses 1-7.

54

Page 62: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Another potential moderator of the promotion-performance relationships stated

in this research is the product’s purchase frequency. For instance, razorblades are

purchased more frequently than toasters. In either situation, this often corresponds

with, but yet is distinct from, the product durability. It is possible for durable products

(e.g., toasters) to become commodities. It is proposed here that the perceived

consumability of these products is affected by price promotions. No evidence is found

in the literature to propose that the preceding hypotheses will be different (e.g.,

positive effect, negative effect) for more frequent (i.e., health and beauty aid) products

vs. less frequent (i.e., small appliance) products. However, will the effect size be

similar, or will the effect size of one type of product be larger? Thus,

RQ2: Is the effect of price promotions in H1 to H7 greater for more frequently purchased products than for less frequently purchased products?

Last, this research project investigates when (i.e., under what competitive

conditions) customers are “cherry picking” shopping due to the price promotion versus

“self-indulging” or “hedonic” shopping due to the price promotion. That is, under

what conditions (i.e., competitive intensity, customer demographics) is it more likely

that consumers will respond to the price promotion by purchasing more reduced price

products—reducing profitability growth versus sales growth (i.e., a reduction in

profitability margin)? Alternatively, under what conditions is it more likely that

consumers will respond to the price promotions by purchasing more regularly priced

products—increasing profitability growth versus sales growth (i.e., an increase in

profitability margin)? Both situations are consistent with the theory of expandable

55

Page 63: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

purchasing and consumption. However, there is no research published (based on an in-

depth review of the literature) to indicate when one effect over the other effect will be

found in the data analysis of this research study. Thus, the investigation takes a

research question form:

RQ3a: When does the moderating effect of H8a to H14k decrease profitability margin for either the product or market basket?

RQ3b: When does the moderating effect of H8a to H14k increase profitability

margin for either the product or market basket?

56

Page 64: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

CHAPTER 3

RESEARCH DESIGN

Samples

The samples for this research consists of weekly, point-of-sale, stock-keeping

unit (hereafter, SKU), scanner data for manufacturer-branded products collected in all

of the approximately 450 stores of an every-day-low-price, general-merchandise

retailer operating in the northeastern and midwestern United States. Several reasons

exist as to why this sample kind is appropriate for the study of retail price promotions.

First, retailer store-level scanner data provides a way to track purchase

behavior that is (1) store specific and (2) more likely to be error free than either

household scanner data or surveys of customers. According to Bucklin and Gupta

(1999), store level scanner data is less prone to sample selection bias than household

level (or survey) scanner (panel) data. Further, as indicated by Sriram, Balachander,

and Kalwani (2007, p. 61), because “store-level data can be obtained across several

retailers in various geographic regions, managers can use these data to track the health

of their brands across geographic regions.” This is important because, in the words of

Kamakura and Kand (2007, p. 160), “while managers of retail chains develop price-

promotions policies that are consistent with the marketing strategy at the chain level,

they should implement theses policies in a way that is most effective at each store.”

Second, demand for manufacturer-branded products has been shown to be

more price-elastic than demand for non-branded or private-label products (e.g.,

57

Page 65: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Sethuraman 1995). For example, Sethuraman (1995), finds that “national brands with

large market share have significant influence on private-label sales but they are less

likely to be affected by private-label price cuts.” The increased probabilty that the

demand for branded products will be more elastic increases the opportunity to

examine hypotheses regarding market basket, shareholder inventory investment, and

moderator effects.

Third, a general-merchandise retail store carries a greater variety in product

categories than other retailer type stores modeled in the literature (e.g., grocery stores,

drug stores). The increased variety makes it “easier for consumers to combine multiple

visits to multiple stores” (Dellaert, et al. 1998, p. 177). Indeed, customers can

normally combine apparel, home (e.g., furniture, bed and bath, kitchen), health and

beauty aids, grocery, and pharmacy product purchases at a general merchandise retail

store (e.g., Big Lots, B.J.’s, Costco, Dollar General, K-Mart, Shopko, Target, Wal-

Mart). The increased quantity/variety of categories permit greater investigation into

the effects of price promotions on customer market basket choices.

Fourth, as indicated by Hock, Dreze, and Purk (1994, p. 16) “the ‘every day

low price’ (EDLP) format…has experienced rapid growth and media popularity.” The

increase in customer acceptance and trust in EDLP claims permits a more transparent

view of the effectiveness of price discounts by decreasing the potential for customers

doubting price claims. Price discount claims are made by low-consistency cues (e.g.,

“was” or “regularly priced” on the shelf product-label). See Grewal, Marmorstein, and

Sharma (1996) for more discussion on low-consistency cues. For example, a “was

58

Page 66: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

$4.00” in a “was $4.00, now $3.00” claim in an EDLP store is likely to be more

trusted by customers versus claims in other stores using other pricing strategies (where

customers would be more likely to raise doubts as to the motives behind the “was

$4.00.”) In contrast, to what extent does the customer percieve that the non-EDLP

store is inflating the price to make the “now” price more attractive? This perception

seems less likely to occur by customers in EDLP stores (e.g., Fishman 2006; Hock,

Dreze, and Purk 1994). Also, price promotions are less frequent in EDLP stores.

Blattburg, Briesch, and Fox (1995, p. 124), citing Bolton (1989), Raju (1992), Winer

(1986), and Putler (1992), state “the greater the frequency of deals, the lower the

height of the deal spike.” Thus, the potential effects of price promotions should be

more evident in EDLP stores.

Fifth, Kamakura and Kang (2007, p. 161) call for price promotion research that

uses more than “a single store or small set of competing stores (e.g., two stores)” retail

store locations. Indeed, the approximately 450 store sample for each product SKU in

this research study permits investigation into the “possibility that each store in a retail

chain serves a distinctive trade areas responding different to price promotions,” and

does not “restrict cross-elasticities to be that same for all stores” (Kamakura and Kang

2007, p. 161). These several reasons provide description of why this sample kind is

appropriate for the study of retail price promotions. In addition to the preceding

reasons, several filters are applied to potential product SKU’s to increase the

appropriateness of the research setting for investigating the effects of price promotions

on retailer profitability.

59

Page 67: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Data Collection

Several filters were applied to available product SKUs (that were price-

promoted by the retailer) to arrive at the final sample that is consistent with the stated

research purpose to investigate the impact of retailer price promotions on retailer

performance.

First, and central to the stated research purpose, product samples are limited to

particular products where pure price promotions (i.e., no additional promotional

activities for the investigated products) were taken at some point during the period

June 2004 to July 2006. By using only pure price promotions, the first instance of

general customer price-reduction awareness is in-store when the customer views the

shelf label (which states the prior and current price, both in dollars). This setting

contains the advantaging of preventing a needed simultaneous investigation of

customer price knowledge and the source of customer price comparison. Customers

can refer to different references when making price comparisons, including external

(e.g., shelf label, advertisement) or internal (i.e., memory based) references. Because

this research controls for advertising (through exclusion), all customers first see the

price change at the shelf. Thus, we do not need to investigate their ability to recall and

advertised price promotion (because there is no advertising involved), etc. Rather, all

customers are exposed to and using the same price comparison (i.e., the shelf label).

For additional discussion on the different types of external price references, see

Vanhuele and Dreze (2002).

60

Page 68: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Second, following the finding of Ailawadi et al. (2006) that the greatest

variance (in their study) is in the general merchandise and health and beauty

merchandise categories, the product SKUs in this study are chosen from these same

merchandise categories. In contrast to both the drug store setting of Ailawadi et al

(2006) and the grocery setting of other, related research (e.g., Janakiraman, Meyer,

Morales 2006; Mulhern and Leone 1991; Walters 1991; Van Heerde, Leeflang, and

Wittink 2004), however, this research uses samples drawn from an every-day-low

price, general-merchandise retailer in the United States, which permits investigation of

the generalization of their findings (to a general-merchandise store context), in

addition to the new contributions on total shopping cart consumption, shareholder

return, and competitive intensity.

Third, products are selected for which the product market baskets do not

include any of the other products being investigated in the study (i.e., unique market

baskets). Unique market baskets are operationalized here as market baskets in which

the overlap (in investigated products) is less than 1 in a 100 common occurrences.

This filter was accomplished through analyzing system generated reports of the most

commonly occurring items in market baskets for each potential price-promoted

product over the course of a year. It should be noted that while the filters decrease the

level of “noise” in the sample data, there is always the possibility that some other

products may be price-promoted that customers bought as part of some of the market

baskets because of the natural field experimental setting of this research. This does not

imply, however, that the findings of this research are overstated. Rather, it suggests

61

Page 69: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

that any corroboration of hypotheses in the data analysis is understated (because the

relationships are significant even in the presence of the potential marketing mix

activities on other products in the market baskets).

In all, four product SKU samples, each of N = approximately 450 (same) retail

stores, are found that match the pure promotion, timeframe, merchandise category, and

unique market basket filters. The data were made available by the respective products’

manufacturers. Vice presidents of marketing or sales departments for manufacturing

firms identified product SKUs that match all of the specified filter criteria. The

manufacturer contacts then queried the data for the variables described in the

measurement section from the retailer scanner database for each product SKU

matching the filter constraints. Per their request, the products, brands, and retailer

names are not disclosed. Two of the product SKU samples are frequently purchased

products from the health and beauty aid category. The other two of the product SKU

samples are more infrequently purchased products from the small appliances category.

Measurement of Variables

All of the construct are measured at the product-store level. All pricing,

product performance, and market basket variables are measured for each product by

retail store, aggregated at a weekly level. These data come directly from the retail

scanners in each store location that is maintained is a centralized, corporate database.

All competitive intensity and customer demographic variables are, likewise, calculated

for each store, as described next.

62

Page 70: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Product Sales

Product Sales is measured as net sales dollars. Net sales dollars is calculated as

the total sales dollars, minus return dollar figures, retained by the store (Risch, 1987,

p. 116). According to Walters and MacKenzie (1988, p. 56), “Using dollars instead of

units is more managerially relevant because retailers measure the success or failure of

their decisions, including the implementation of promotional activities, in dollars and

not in units.” It is aggregated at the weekly level by store.

Product Profitability

Product profitability is measured as gross margin dollars. Gross margin is

calculated as the difference between net sales and the billed cost price (before shrink

and allowances) of the merchandise sold, taking into account any reductions due to

sales discounts plus markdowns (e.g., Ailawadi et al. 2006). This definition of gross

margin is closely related to the definition of Risch (1987, p. 296) in which gross

margin is equal to the maintained markon minus the cost of alteration and cash

discounts. The two approaches only differ on the inclusion of cash discounts (as there

are no alteration costs for the general merchandise products studied here). The more

commonly used definition, as adopted here, excludes cash discounts (placing them in

gross profit or maintained profit), and is more often used by practitioners (e.g., Dunne

and Lusch 2006).

Gross profit, in comparison, is equal to gross margin minus operating expenses

(see Risch 1987, p. 290), and it is not calculated here. While the cash discount

component of gross profit can be calculated (combining available information on the

63

Page 71: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

inventory turnover and the retailer’s accounting method [e.g., LIFO]), the allocation of

operating expenses across products is not known. Indeed, the retailer has never

attempted to allocate it across its products. Given that there is no consensus on how to

allocate operating expenses (e.g., how much does a can of tuna really cost? Or, how

much store labor is allocated to a can of tuna?), gross margin preferred to gross profit

(due to avoidance of product-operating expense allocations). The gross margin is

aggregated at the weekly level by store.

Product Inventory Turnover

Product inventory turnover is measured as net sales dollars divided by average

inventory dollars at billed cost to the retailer (Risch 1987, p. 322). Average inventory

dollars are calculated in this research study as the averages of beginning inventory (at

billed cost dollars) and ending inventory (at billed cost dollars) for a given week and

then annualizing the figure. While average inventory could be computed for any

period of time, it is calculated here by week to be consistent with the other metrics,

permitting other calculations (e.g., GMROII, GMROSI).

Market Basket Sales

Market basket sales is measured as the aggregated net sales dollars for all of

the products acquired by a customer during a single purchase event at the store register

(i.e., the total shopping cart purchased). The calculation is the same as the product net

sales dollars calculation. The measure is aggregated at the weekly level by store for

market baskets that contained the product SKU under investigation.

64

Page 72: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Market Basket Profitability

Market basket profitability is measured as the aggregated gross margin dollars

for all of the products acquired by a customer during a single purchase event at the

store register (i.e., the total shopping cart purchased). The calculation is the same as

the product gross margin dollars calculation. It is aggregated at the weekly level by

store for market baskets that contained the product SKU under investigation.

Market Basket Item Count

Market basket item count is measured as the number of products in a particular

market basket. It is a proxy for market basket inventory turnover—which would be

nearly impossible to calculate (i.e., it would require separate reports on every item that

is located in any of the market baskets in any store during any week for the price-

promoted product SKUs). The number of system queries would be astronomical. The

number of items in the market basket, in contrast, is calculable, and gives a rough

estimation of whether the store is selling more products per period of time (increasing

overall inventory turnover). The market basket item count is aggregated at the weekly

level by store for market baskets that contained the product SKU under investigation.

Gross Margin Return on Inventory Investment (GMROII)

GMROII is a calculation—shown in Chapter 2 on page 42. It combines

inventory turnover and gross margin percentage (each defined earlier).

Gross Margin Return on Shareholder Investment (GMROSI)

GMROSI is a calculation—shown in Chapter 2 on page 46. It combines

inventory turnover, gross margin percentage, and adjusted product cost. The adjusted

65

Page 73: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

product cost reflects the internal rate of return. Several potential internal rates of

return from the response surface are used in calculating potential GMROSI scenarios.

Store Assortment

Store assortment is measured through a proxy, the store’s actual physical size

(e.g., 122,000 square feet), because the number of product combinations at the store

level is astronomical, making the use of all product combinations beyond available

degrees of freedom for statistical testing. However, prior literature has shown store

size to be an adequate proxy for store assortment (e.g., Kamakura and Kang 2007), as

larger stores do, typically, carry additional breadth and depth of products, increasing

the assortment.

Competitive Intensity

Competitor stores of 30 potential, major competitors are geographically

identified for each store in the product SKU samples, first by census data, and then

confirmed (as actual store competitors) by the store manager. This store-level

competitor information is documented and retrieved from the scanner database. Some

of the competitors compete in particular categories, while others compete across many

categories. Some of the competitors carry higher priced products, while others carry

lower priced products. This research will account for these complexities using R-A

theory as a foundation. Adopting this foundation, the research will distinguish between

category and store competitors in modeling the complexities by (1) using indicator

(i.e., dummy) variables for each of the 30 competitor and (2) identifying whether a

given competitor is a total store competitor or a category competitor (as indicated by

66

Page 74: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

the retailer and confirmed through a expert panel survey conducted by the author). The

increased sensitivity of these tests (over prior work) will advance understanding of the

moderating effects of competitive intensity on the relationship between retailer

promotional activities and retailer performance.

Customer Education Level

Customer education is reported at the store level. It is computed by the retailer

using census and other third party data for the store selling area and adjusted, as

deemed appropriate, by the store manager. Percentages are given for the following

categories: some high school, high school, some college.

Customer Age

Customer age is reported at the store level. It is computed by the retailer using

census and other third party data for the store selling area and adjusted, as deemed

appropriate, by the store manager. Percentages (of total customer population for each

store) are given for each of the following categories: 18 to 34, 35 to 54, 55 to 64, 65

and older.

Customer Ethnicity

Customer ethnicity is reported at the store level. It is computed by the retailer

using census and other third party data for the store selling area and adjusted, as

deemed appropriate, by the store manager. Percentages (of total customer population

for each store) are given for each of the following ethnicities: Caucasian, African

American, Latin.

67

Page 75: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Customer Income Level

Customer income level is reported at the store level. It is computed by the

retailer using census and other third party data for the store selling area and adjusted,

as deemed appropriate, by the store manager. Percentages (of total customer

population for each store) are given for each of the following household income

ranges: under $30,000; $30,000 to $49,999; $50,000 to $99,999; $100,000 or more.

68

Page 76: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

CHAPTER 4

DATA ANALYSIS

In all, four product SKU samples, each of N = approximately 450 (same) retail

stores, are found that match the pure promotion, timeframe, merchandise category, and

unique market basket filters. Samples A and B are frequently purchased products from

the health and beauty aid category. Samples C and D are less frequently purchased

products from the small appliances category. Sample A is a loss leader. That is, the

item was price promoted at a retail price below its cost to the retailer (known from

examination of the cost and price information). There is no evidence that customers

were generally aware of this (as markup information is confidential), and so while it

changes the interpretation of the profitability and cash flow analysis for the product, it

should not impact consumer behavior in a manner different from the other products.

That is, customers only saw the differences in retail prices (e.g., was/now) on the shelf

labels.

The price promotion’s effect on performance is measured against two baselines

for each product SKU sample. The first baseline, following Abraham and Lodish

(1993), Ailawadi et al (2006), and Kopalle et al. (1999), consists of a moving average

baseline. Because sufficient prior data is available, a second baseline, consisting of

prior-year, same-period data, is also used. The second baseline is more consistent with

practice. That is, firms publicly report comparisons to prior year, same time period

data. Moreover, they typically do not report comparisons to the prior month or quarter.

69

Page 77: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Paired samples T-test procedures are used to compare the difference between each

baseline, in turn, and the customer purchase behavior after the price promotion occurs.

Data Analysis of H1 to H5: Product and Market Basket Performance

Table 2 summarizes the paired T-test comparison results of differences in

product sales, product inventory turnover, and product gross margin for each of the

four samples. The data provide strong support for hypotheses H1 and H2. That is, (1)

pure price promotions of a product do have a positive effect on that product’s dollar

sales (at the product-store level), and (2) pure price promotions of a product do have a

positive effect on that product’s inventory turnover (at the product-store level). Also,

with the understandable exception of the loss leader sample, pure price promotions of

a product do have a positive effect on that product’s total dollar profitability for the

analyzed time frame (at the product-store level). That is, the aggregated, incremental

increase in gross profit dollars more than offsets the per item decrease in gross profit

dollars (that happens because the item is price reduced, decreasing gross margin per

unit).

Table 3 summarizes the paired T-test comparison results of differences in

market basket variables. The data provide strong support for hypotheses H3, H4a,

H4b, and H5. That is, (1) pure price promotions of a product do have positive effects

on market basket sales, (2) pure price promotions of a product do have positive effects

on the number of market baskets (i.e., an increase in the number of people buying the

product—rather than just the same number of people buying more of the product), (3)

pure price promotions of a product do have positive effects on market basket item

70

Page 78: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

count (i.e., number of products in the market basket), and perhaps most importantly to

overall store and company performance (4) pure price promotions of a product do

have positive effects on market basket profitability. Indeed, the increase in market

basket profitability more than offsets the loss in item profitability for the loss leader

sample store group.

Data Analysis of H6 to H7: GMROII and GMROSI

In regards to comparing changes in GMROII and GMROSI, first, a simulated

response surface is constructed to provide a theoretical range of GMROII and

GMROSI values that retailers, in general, might expect to achieve under different

variable (i.e., gross margin, inventory turnover, contracted terms of payment, and

internal rate of return) combinations. See Myers (1971) and Myers and Montgomery

(1995) for overviews of response surface methodology.

The response surface methodology (RSM) has been used in marketing

research, for instance, to model interdependence in supplier-distributor channel

relationships (Kim and Hsieh 2003), customer service satisfaction (Danaher 1997),

and, more related to this research, congruence between pricing strategies and venture

strategies (Myers 2004). While response surfaces can investigate the relationship

between multiple independent variables and a response (i.e., dependent) variable (Box

and Wilson 1951), the RSM has occasionally been criticized because the optimization

is typically done in a model where the coefficients are hypothetical estimations, rather

than real-world measurements. In this research, the model coefficients are not

71

Page 79: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

estimations. Rather, the coefficients are data values from quarterly and annual reports

of the thirty publicly-held retailers in CNUM 5331 (e.g., Dollar Tree, Target, Wal-

Mart), 5411 (e.g., Kroger, Safeway), 5912 (e.g., CVS, Rite Aid, Walgreens), and 5399

(e.g.., BJ’s, Costco) to establish parameter boundaries and scales. These thirty firms

are selected because they are also the same thirty competitors being included in the

competitive intensity analysis. Specifically, ranges for the inventory turnover, gross

margin, account payable (terms of payment), and internal rate of return variables are

developed based on analyses of the retailers’ quarterly and annual reports. Using the

public report data permits the advantages of RSM while avoiding potential criticism of

common estimation practices (of how one defines the theoretical ranges).

One significant advantage of the response surface methodology is that it

permits investigation of the cost of capital while avoiding entry into the debate over its

calculation (see Chen, Dhaliwal, and Xie 2006). Instead, practitioners can identify

their organizations’ cost of capital (as they compute it) and then find the "path of

steepest ascent" in the direction of maximum GMROII or GMROSI on the response

surface (which indicates the best "direction" for management to incorporate into their

pricing and inventory strategies).

The response surface analysis indicates quadratic modeling for both the

“GMROII” and “Net Change in Market Basket Profitability” models. Canonical

analysis shows that the eigenvectors’ parameters are all positive for both models,

indicating directions of upward curvature shape in the response surfaces. However,

there is no apparent single optimum. See Figure 2.

72

Page 80: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

<< Insert Figure 2 About Here >>

As a result, ridge regression analysis was performed to determine what

direction should be searched on the grid to locate a superior gain. Ridge regression

reduces the standard errors by adding a degree of bias to the regression estimates. The

superior k value (bias addition) to the ridge regression is k = 0.019564. The resulting

ridge regression parameter estimates are found in Table 4

<< Insert Table 4 About Here >>

GMROII and GMROSI metrics are calculated (using the equations in Chapter

2 on pages 36 and 40) for the post promotion period and both the first (i.e., same year,

prior period) and second (i.e., prior year, same period data) baselines, in turn, for each

of the four natural field experiments. Paired sample T-tests are used to compare pre-

promotion and post-promotion GMROII and GMROSI figures. These results are then

discussed in terms of their location on the generated response surface.

Table 5 summarizes the paired T-test comparison results of differences in

GMROII and GMROSI variables. The data provide strong support for hypotheses H6

and H7. That is, (1) pure price promotions of products, in a majority of cases, have

positive effects on promoted products’ GMROII, and (2) pure price promotions of

products, in a majority of cases, have positive effects on promoted products’

GMROSI.

<< Insert Table 5 About Here >>

73

Page 81: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

The GMROII results from Table 5 are mapped on the response surface in

Figure 2, resulting in Figure 3. As seen in Figure 3, the locations of pre-promotional

period values are in the normal range of response surface values. Indeed, they appear

to be in the “Big Middle” where most values are located. The promotional timeframe

data points are all located along ascension paths on or above the surface. Thus, the

impact of price promotions on the gross margin percentage and the average inventory

turnover is such that it positively impacts the GMROII in all cases investigated.

< Insert Figure 3 about Here >>

Comparing the GMROII and GMROSI figures for the eight comparisons, the

difference between the two metrics (i.e., GMROSI change minus the GMROII

change) results in an equal interpretation in three of the eight comparisons and a

stronger confirmation in five of the eight comparisons (where the difference was .1, .1,

.1, .2, an .4). Note that a difference of .1 is interpreted as an additional 10% return on a

dollar invested in inventory. Thus, a range of zero to .4 suggests that GMROII does, in

all explored cases, underestimate the impact of marketing activities on shareholder

investment.

Analysis of Hypotheses 8a to RQ3b: Moderating Variables

In regards to the hypotheses related to competitive intensity and customer

demographics, correlations are constructed for the potential moderator variables and

the net change in values of the variables from hypotheses 3 through 7, consistent with

prior literature (e.g., Ailawadi et al. 2006, p. 528-9).

74

Page 82: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Table 6 presents the correlates of the potential moderating variables on the net

market basket performance (sales, profits, inventory turnover) and net product

GMROII and GMROSI data. The results indicate that increases in merchandise

heterogeneity strengthen the relationships in hypotheses 3-5 on market basket

performance, consistent with the moderator effects proposed in hypotheses 8, but

weaken the relationships in H6 and H7 on GMROII and GMROSI.

In regards to the potential roles of competitive intensity, the results indicate

that increases in the number of total competitors strengthens the impact of the pure

price promotion on the gap between current market basket sales and prior year (same

period) market basket sales; they support moderator effects proposed in hypotheses 9.

Interestingly, though, an effect was not statistically significant for the number of

primary competitors or for the number of secondary competitors (i.e., hypotheses 10

and 11). One feasible explanation is that the effect of competition occurs more at a

local level (versus the analysis here spanning several states). That is, aggregating only

a particular set of category competitors and a separate set of other competitors does

not account for the fact that not all competitors are present in all markets, and thus

either the total number of competitors should be used, or the data should be modeled

at a more disaggregate level.

As to the moderating effects of consumer demographics, increases in the level

of customer education do have a positive effect on three of six of the relationships in

hypotheses 3-7. Increases in the age of customers do have a negative effect on five of

six of the relationships in hypotheses 3-7 (with the other effect being not statistically

75

Page 83: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

significant). Also, increases in the amount of customer income do have a positive

effect on the relationships in hypotheses 6 and 7. The results on market basket net

performance are mixed. Thus, mixed support is found for the proposed moderating

effects in hypotheses 13. Likewise, increases in the percentages of African American

customers do not have a statistically significant effect on the relationships in

hypotheses 1-7; support is not found for hypotheses 14. However, increases in the

percentages of Hispanic customers do have a positive effect on five of six of the

relationships in hypotheses 1-7 (with the other effect being not statistically

significant). Thus, support is found for hypotheses 15.

As to the second, general research question, the data show that the effect of

price promotions in H1 through H7 are greater for more frequently purchases products

(i.e., products A and B) than for less frequently purchased products (i.e., products C

and D).

As to the third, general research question, the data show that the moderating

effect of H8a to H14k increase the profitability margin for the market basket when there

is increased (1) merchandise heterogeneity, (2) competitors operating stores, (3)

household size, and (4) percentage of customers who are Latin American.

76

Page 84: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

CHAPTER 5

DISCUSSION AND CONCLUSION

Discussion

Prior literature suggests that price promotions are not beneficial because, at

least in part, the increased sales from the promotion are the result of product switching

or stockpiling. Such arguments are grounded in metaphors and logic that are derived

from an equilibrium neoclassical economic research tradition. Adopting a dynamic

competition perspective based on resource advantage theory, this research study

proposes that price promotions can be beneficial to companies who market and sell

products to consumers. The phrase “who market” is not equivalent to the phrase “who

sell.” Marketing involves pricing, advertising, promotion, and so forth. Selling does

not, by definition, necessarily involve these elements. The phrase “who market and

sell products to consumers” addresses the generalizability of this research study.

Indeed, it both includes and excludes certain manufacturers and retailers. Some

manufacturers such as Nike or Sony both market and sell directly to consumers

through their own consumer retail outlet stores or online. In contrast, a few small

retailers may not have enough relational control in their purchasing relationships to

change prices. Thus they may sell, but do not market (e.g., they cannot change or

control marketing mix activities such as pricing). Thus, the findings of this study apply

to most retailers as well as to a number of manufacturers. Therefore, some

manufacturers could be included (or substituted) in the phrases on retailers

performance.

77

Page 85: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

In particular, this research proposes that price promotions have a positive

effect on (1) the retailer’s total market basket performance (i.e., market basket sales,

profitability, and inventory turnover), and (2) the price-promoted products’ financial

return on shareholder investment in product-level inventory.

The analysis of four unique samples, each comprised of product and related

market basket and moderator data for 450 stores, supports the hypotheses. That is,

consumers, on average, do purchase more of the product and of other items. See

Figures 4a to 4d. The result is that while product gross margin percent decreases,

product gross margin dollars increases in each of the four samples—an effectiveness

gain. Furthermore, the decrease in margin percent is offset by the inventory

acceleration that results in a positive gain on the profit generated per dollar invested

by companies’ shareholders—an efficiency gain. Also, the market basket profit gains

are much greater than the product profitability decreases. Moreover, the GMROII and

GMROSI metrics indicate a positive gain. Thus, product price promotions can be

beneficial to retailers and to the retailers’ shareholders.

<< Insert Figure 4a to 4d about here >>

Implications for Practice

The results indicate that, at least for the 450 retail stores investigated here, pure

price promotions—simply reducing the price of the item without any advertising or

displays—can result in changed consumer purchasing behavior. That is, consumers as

a whole both purchase more often (e.g., the number of market basket carts) and they

purchase more (e.g., market basket sales dollars). Indeed, the results indicate that pure

78

Page 86: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

price promotions can provide both an effectiveness (e.g., product and market basket

sales and profit dollar performance) and an efficiency (e.g., inventory turnover,

GMROII, GMROSI) advantage to the retailers. Indeed, the gain of market basket

profit more than offsets the loss of product profit—even for the analyzed loss leaders

in Sample A.

The results provide reasons for retailers to be strategic in their approach to

price promotions, and to use care in how they measure their impact, both in the metric

and the baseline. As espoused in R-A theory, “All strategies (at the business-unit

level) involve, at the minimum, the identification of (1) market segments, (2)

appropriate market offerings, and (3) the resources required to produce the offerings”

(Hunt 2000, p. 131). For instance, given that some customer segments will cherry pick

items while other segments will expand their purchasing, merchandise assortments and

other activities could be customized (e.g., the identification and making of

“appropriate” market offerings) for identified market segments that participate more in

expandable purchasing activities, resulting in firm superior financial performance.

Part of the strategy could also include determination of the implications for logistics

(flow through constraints) and operations (potential out of stocks or merchandise

presentation effects). Care should be taken to distinguish between whether

observations that customers report having had to park farther away from the stores,

more merchandise offerings are out-of-stock, the restrooms are more dirty, and the

check-out lines are longer are (as they are commonly assumed to be) indeed evidences

of poor management or, in contrast, are simply (and to some extent uncontrollably) the

79

Page 87: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

result of increased inventory turnover. Indeed, retailers should take care in how much

they increase the patronage of customers through their pricing strategies so as not to

reach or exceed a “critical mass” of customer traffic within a store. Increased

“shopping momentum” can be good (see Dhar. Huber, and Khan 2007)—until it

approaches or exceeds critical mass. Price promotions, as shown here, increase

inventory turnover as more customers buy and customers buy more. Increased

customer patronage can affect the store atmospherics, having unintended

consequences on many customers.

Limitations and Implications for Future Research

In regards to the marketing literature on promotions, while this study provides

a unique, important first view of the effects of pure, retail price promotions on

consumer behavior at the market basket level, in so doing it excludes investigation of

the effects of promotions where the products are retailer advertised but not retailer

price discounted (e.g., to create awareness of the product availability at the store) or

retailer advertised and retailer priced discounted (e.g., to increase customers visiting

the store). Thus, future research is needed that investigates these other promotional

activities, including comparisons across different types of promotions.

As the study shows that there is, indeed, an effect from pure price promotions

on consumer purchasing behavior, future research is needed that investigates how

consumers mentally make such changes. To what extent is it due to shopping

momentum (see Dhar. Huber, and Khan 2007), hedonic enjoyment, or utilitarian

80

Page 88: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

savings? Or, recognizing differences in customer segment profitability, for which

customers segments might it be more of one of these or other motivations?

Possibly one of the most important areas for future research is investigation of

promotional competition at the local (geographical) level. This research has advanced

knowledge on how competition can affect consumers’ response to pure price

promotions. The results of the four sample analyses indicate that the total competitive

intensity has a positive effect on marketing mix activity effectiveness and efficiency

for the 450 stores of a major retail firm. This finding can be explained through

resource advantage theory (as described in Chapter 2). Equilibrium-based economic

theory cannot account for the results. Indeed, equilibrium-based economic theory

indicates that there should be a negative (or opposite) effect. Thus, additional research

is needed that evaluates how different configurations of marketplace competitive

intensity affect consumers’ buying behavior. That is, to what extent are there result

differences between stores where there are no identified local competitors versus

stores facing limited competitive intensity versus stores facing extensive competitive

intensity? Further, is the effect difference negative (in support of equilibrium-based

economic theory) or is it positive (in support of dynamic competition theory)?

Moreover, if competition does have a positive impact on the firm’s price promotions

effects (as in the current study), is it due more to competitors that are more similar to

the firm (i.e., Porter 5-Forces rivalry) or is it due to all competitors (i.e., R-A theory

rivalry)? Such research will provide confirmation as to which theories of competition

best represent actual consumer purchasing behavior, as the results here provide initial

81

Page 89: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

support that dynamic competition theory is better able to explain the moderating,

positive impact of competitive intensity on the price promotions effects on retailer

financial performance.

82

Page 90: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Table 1. Literature Related to the Effects of Price Promotions on Market Basket Performance

Intra-Category Performance

Inter-Category Performance

Total Market Basket Performance

Store-Level Scanner Data

Van Heerde Leeflang, Wittink (2004, p. 326) *; Van Heerde, Gupta, Wittink (2003, p. 486); Kamakura and Kang (2006, p. 163)

Mulhern and Leone (1991, p. 67) 2 cat; Walters (1991, p.20) 2 pair of 2 cat; Kamakura and Kang (2006, p. 163) 2 cat

This

research study *

Household-Level Scanner Data

Nijs et al. (2001, p. 7-8); Bell, Chiang, Padmanabhan (2002, p. 514)

Chintagunta and Haldar (1998, p. 44) 5 pair of 2 cat

Ailawadi et al. (2006, p. 520)

Survey Shoppers

Van Trijp et al. (1996, p. 285)

Mulhern and Padgett (1995, p. 85)

Experimental (students)

Janakiraman, Meyer, Morales (2006, p. 363), 12 categories *

Read: store-level scanner data is acquired directly from retail point of purchase (POS) scanning of products at checkout in stores. Household-level scanner data is acquired from panels of customers self-reporting purchasing behavior (post-purchase). Total Market Basket Performance is analysis of the entire shopping cart purchased by a customer. Inter-Category Performance refers to analysis of products in different categories—usually two complementary categories (e.g., cake and frosting). Intra-Category refers to analysis of products within the same product categories (e.g., multiple products or brands of cookies). The two prior studies that are asterisked (*) included some products where only pure price promotions were present. All other studies investigated price promotions where advertising and other sales promotion activities were present.

83

Page 91: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Table 2. The Effects of Pure Price Promotions on Product Performance Hypothesis Baseline Sample µ-Chg p H1 (Product Sales) Prior Yr, Same Time A 18.1 < 0.001 H1 Prior Yr, Same Time B 65.6 < 0.001 H1 Prior Yr, Same Time C 4.8 < 0.001 H1 Prior Yr, Same Time D 5.9 < 0.001 H1 Same Yr, Prior Time A 20.1 < 0.001 H1 Same Yr, Prior Time B 104.9 < 0.001 H1 Same Yr, Prior Time C 8.4 < 0.001 H1 Same Yr, Prior Time D 10.2 < 0.001 H2 (Inv Turnover) Prior Yr, Same Time A 4.7 < 0.001 H2 Prior Yr, Same Time B 5.3 < 0.001 H2 Prior Yr, Same Time C 2.3 < 0.001 H2 Prior Yr, Same Time D 2.9 < 0.001 H2 Same Yr, Prior Time A 5.9 < 0.001 H2 Same Yr, Prior Time B 7.1 < 0.001 H2 Same Yr, Prior Time C 3.5 < 0.001 H2 Same Yr, Prior Time D 4.4 < 0.001 RQ1 (Product Profit) Prior Yr, Same Time A (12.4) < 0.001 RQ1 Prior Yr, Same Time B 8.9 < 0.001 RQ1 Prior Yr, Same Time C 0.4 < 0.001 RQ1 Prior Yr, Same Time D 0.5 < 0.001 RQ1 Same Yr, Prior Time A (12.2) < 0.001 RQ1 Same Yr, Prior Time B 28.4 < 0.001 RQ1 Same Yr, Prior Time C 1.8 < 0.001 RQ1 Same Yr, Prior Time D 2.0 < 0.001 Read: Sample size = 450 stores for each sample. Sample A was the loss leader. µ-Chg is measured in retail sales dollars for H1, in annualized inventory turns for H2, in retail gross margin profit dollars for RQ1.

84

Page 92: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Table 3. The Effects of Pure Price Promotions on Market Basket Performance Hypothesis Baseline Sample µ-Chg p H3 (MB Sales) Prior Yr, Same Time A 746.6 < 0.001 H3 Prior Yr, Same Time B 627.6 < 0.001 H3 Prior Yr, Same Time C 197.9 < 0.001 H3 Prior Yr, Same Time D 229.6 < 0.001 H3 Same Yr, Prior Time A 665.7 < 0.001 H3 Same Yr, Prior Time B 854.4 < 0.001 H3 Same Yr, Prior Time C 296.6 < 0.001 H3 Same Yr, Prior Time D 369.0 < 0.001 H4a (# Baskets) Prior Yr, Same Time A 7.5 < 0.001 H4a Prior Yr, Same Time B 4.8 < 0.001 H4a Prior Yr, Same Time C 0.9 < 0.001 H4a Prior Yr, Same Time D 1.6 < 0.001 H4a Same Yr, Prior Time A (7.7) < 0.001 H4a Same Yr, Prior Time B (6.9) < 0.001 H4a Same Yr, Prior Time C (2.6) < 0.001 H4a Same Yr, Prior Time D (3.5) < 0.001 H4b (Item count) Prior Yr, Same Time A 210.0 < 0.001 H4b Prior Yr, Same Time B 77.8 < 0.001 H4b Prior Yr, Same Time C 49.6 < 0.001 H4b Prior Yr, Same Time D 58.7 < 0.001 H4b Same Yr, Prior Time A 197.0 < 0.001 H4b Same Yr, Prior Time B 95.0 < 0.001 H4b Same Yr, Prior Time C 83.5 < 0.001 H4b Same Yr, Prior Time D 102.4 < 0.001 H5 (MB Profit $) Prior Yr, Same Time A 172.4 < 0.001 H5 Prior Yr, Same Time B 84.6 < 0.001 H5 Prior Yr, Same Time C 32.8 < 0.001 H5 Prior Yr, Same Time D 43.6 < 0.001 H5 Same Yr, Prior Time A 165.1 < 0.001 H5 Same Yr, Prior Time B 141.5 < 0.001 H5 Same Yr, Prior Time C 58.9 < 0.001 H5 Same Yr, Prior Time D 77.0 < 0.001 Read: Sample size = 450 stores for each sample. Sample A was the loss leader. µ-Chg is measured in total market basket retail sales dollars for H3, the number of occurring market baskets for H4a, the total number of products in the basket for H4b, and the total market basket retail gross margin profit dollars for H5.

85

Page 93: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Table 4. Standardized Ridge Regression Coefficients-GMROII Response Surface

Variable RegressionCoefficient

Standard Error

StandardizedCoefficient VIF

Intercept -0.895 GM 2.903 0.551 0.170 4.773 INV 0.075 0.011 0.238 5.264 GM^2 1.858 1.087 0.058 5.279 INV^2 -0.001 0.001 -0.041 6.964 GM x INV 1.089 0.042 0.807 4.485

Read: The K= 0.019564. The K value search was performed using the Hoerl's (1976) algorithm. GM = Gross Margin Percentage, INV = Inventory Turnover.

86

Page 94: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Table 5. The Effects of Pure Price Promotions on GMROII and GMROSI

Hypothesis Baseline Sample µ-Chg p H6 (GMROII) Prior Yr, Same Time A (2.4) < 0.001 H6 Prior Yr, Same Time B 0.8 < 0.001 H6 Prior Yr, Same Time C 0.9 < 0.001 H6 Prior Yr, Same Time D 1.3 < 0.001 H6 Same Yr, Prior Time A (2.2) < 0.001 H6 Same Yr, Prior Time B 2.5 < 0.001 H6 Same Yr, Prior Time C 1.6 < 0.001 H6 Same Yr, Prior Time D 2.0 < 0.001 H7 (GMROSI) Prior Yr, Same Time A (2.4) < 0.001 H7 Prior Yr, Same Time B 0.8 < 0.001 H7 Prior Yr, Same Time C 1.0 < 0.001 H7 Prior Yr, Same Time D 1.4 < 0.001 H7 Same Yr, Prior Time A (2.2) < 0.001 H7 Same Yr, Prior Time B 2.6 < 0.001 H7 Same Yr, Prior Time C 1.8 < 0.001 H7 Same Yr, Prior Time D 2.4 < 0.001 Read: Sample size = 450 stores for each sample. Sample A was the loss leader. µ-Chg is measured in unit change of the Gross Margin Return on Inventory Investment for H6 (GMROII) and in the unit change in the Gross Margin Return on Shareholder Investment for H7 (GMROSI).

87

Page 95: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

88

Table 6. The Moderating Effects on H3-H7.

Basket Sales

Basket Profit

# Items

# Baskets GMROII GMROSI

Store Assortment 0.24*** 0.17*** 0.16*** 0.28*** -0.07*** -0.07*** Total Competition 0.18*** 0.06*** 0.23*** 0.11*** 0.05** 0.05* Cat Competition n.s. n.s. n.s. n.s. n.s. n.s. Other Competition n.s. n.s. -0.04* -0.09* 0.04* 0.04* Higher Education n.s. n.s. 0.08*** n.s. 0.09*** 0.08*** Age 65+ -0.07*** n.s. -0.09*** -0.05** -0.07*** -0.07*** Household Size 0.07*** 0.05** 0.07** n.s. 0.12*** 0.11*** < $50,000 n.s. n.s. n.s. 0.05** -0.11*** -0.11*** $100,000 + n.s. n.s. 0.05** -0.04* 0.11*** 0.10*** African American n.s. n.s. n.s. n.s. n.s. n.s. Latin 0.08*** 0.04* 0.11*** n.s. 0.06*** 0.06* Read: Market Basket refers to the aggregated volume of the variable (whether it be retail sales dollars, gross profit dollars, total number of items, or the total number of baskets) for all of the products acquired by a customer during a single purchase event at the store register (i.e., the total shopping cart purchased). Gross Margin Return on Inventory Investment (GMROII) is a calculation—shown in Chapter 2 on page 42. It combines inventory turnover and gross margin percentage. Gross Margin Return on Shareholder Investment (GMROSI) is a calculation—shown in Chapter 2 on page 46. It combines inventory turnover, gross margin percentage, and adjusted product cost. Intra-store competition, or store assortment, is measured through a proxy, the store’s actual physical size (e.g., 122,000 square feet). Total Competition Intensity is the aggregation of the number of competitor stores (up to 30 potential, major competitors) that are geographically identified for each store in the product SKU samples, first by census data, and then confirmed (as actual store competitors) by the store manager. Category competition refers to the total number of competitors who compete with the retailer on the product(s) under investigation, but not at the total store (e.g., Target and Staples competing on 3 ring binders). Other Competition refers to stores that compete with firm in other areas, but not for the product(s) under investigation (e.g., Walgreens competes with Best Buy, but not for laptop computers). Higher Customer Education is the percentage of customers who have attended at least some college. Customer Age (Age 65+) is the percentages of total customer population for each store that are age 65 and older. Income Level is the percentages of total customer population for each store with total household income ranges of less than 50,000 or, for the higher end, $100,000 or more. Customer ethnicity is measured as the percentages of total customer population for each store that are African American or Latin.

Page 96: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Figure 1: A Causal Model of the Effects of Retail Price Promotions

Product • Sales • Inventory Turnover• Profitability

Market Basket • Sales • Item Count & Freq.• Profitability

Product Price Promotion

Return Metrics• GMROII• GMROSI

Competitive Intensity• Industry Based• Category Based• Demographic Based

H1 H2 RQ1

H3 to H5

H6 to H7H8a to H14k,

RQ2 to RQ3b

Product • Sales • Inventory Turnover• Profitability

Market Basket • Sales • Item Count & Freq.• Profitability

Product Price Promotion

Return Metrics• GMROII• GMROSI

Competitive Intensity• Industry Based• Category Based• Demographic Based

H1 H2 RQ1

H3 to H5

H6 to H7H8a to H14k,

RQ2 to RQ3b

Read: The research investigates what the effect of retail price promotions is on measures of product performance (H1, H2, RQ1), market basket performance (H3 to H5), and shareholder investment in the retailer (H6 and H7). Further, the research investigates the moderating effects of competitive intensity and consumer demographics (H8a to Hk14). The dashed line from “market basket” to “return metrics” indicates that while this path is not computable in this project’s acquired database, the return metrics are such that they could be computed, if not aggregated, at multiple levels (e.g., product, category, market basket)—a suggestion for future research. Also, the dashed lines from “product” to “market basket” are future research suggestions (that require a different method).

89

Page 97: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Figure 2. Retailers’ GMROII Response Surface Map

Read GMROII is the Gross Margin Return on Inventory Investment. GM is the product’s gross margin percentage. INV is the product’s inventory turnover rate (annualized). The arrow in the graph represents the path of highest accent, or where the most gain in GMROII would occur given a unit gain in either GM or INV. The quadratic response surface indicates that while there is a steep accent as GM and INV rates are changed, there is a point at which additional change results in a declining return.

90

Page 98: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Figure 3. Retailers’ GMROII Response Surface Map with Promotion Data Overlay

Read: The response surface construction is based on Figure 2. GMROII is the Gross Margin Return on Inventory Investment. GM is the product’s gross margin percentage. INV is the product’s inventory turnover rate (annualized). The Circle Points = Sample A, Square Points = Sample B, Triangle Points = Sample C, Diamond Points = Sample D, for the 4 week aggregates for the prior year, same period baseline and the current year, current period promotional period. In all four samples, the grid locations accended the grid to the right and rear. The price promotions of Samples B, C, and D resulted in a decrease in GM and an increase in INV, which when combined resulted in a positive effect on product GMROII. Product A had a negative GMROII after the promotion, which is logical given that it was a loss leader (i.e., price promoted at a retail below cost).

91

Page 99: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

F

92

igure 4a. Response Surface Map of Sample A Market Basket Profitability

Figure 4b. Response Surface Map of Sample A Market Basket Profitability

Page 100: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Figure 4c. Response Surface Map of Sample C Market Basket Profitability

93

Page 101: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

Figure 4d. Response Surface Map of Sample D Market Basket Profitability

94

Page 102: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

95

Page 103: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

96

REFERENCES Abraham, Magid M. and Leonard M. Lodish (1993), “An Implemented System for Improving

Promotion Productivity Using Store Scanner Data,” Marketing Science, 12 (3), 248-69.

Adler, Lee (1963), “Sales Promotion Effectiveness Can Be Measured,” Journal of Marketing,

27(4), 69-70. Ahern, John T., Jr. and Patrick L Romano (1979), “Managing Inventories and Profits Through

GMROI,” Management Accounting, 61 (2), 22. Ailawadi, Kusum L. and Scott A. Neslin (1998), “The Effect of Promotion on Consumption:

Buying More and Consuming It Faster,” Journal of Marketing Research, 35 (August), 390–98.

Ailawadi, Kusum L, Bari A. Harlam, Jacques César, and David Trounce (2006), “Promotion

Profitability for a Retailer: The Role of Promotion, Brand, Category, and Store Characteristics,” Journal of Marketing Research, 43 (4), 518-35.

Anderson, Paul (1979), “The Marketing Management/Finance Interface,” American

Marketing Association Educators’ Conference Proceedings, Neil Beckwith et al., eds. Chicago: American Marketing Association, 325-29.

Assuncao, Joao L. and Robert J. Meyer (1993),”The Rational Effect of Price Promotions on

Sales and Consumption,” Management Science, 39 (5), 517-535. Baker, Malcolm, Richard S. Ruback, and Jeffrey Wurgler (2007), “Behavioral Corporate

Finance: A Survey,” in B. Espen Eckbo (ed), Handbook in Corporate Finance, Volume 1: Empirical Corporate Finance, Elsevier North Holland, forthcoming.

Bell, David R., Chiang, and V. Padmanabhan (1999), “The Decomposition of Promotional

Response: An Empirical Generalization,” Marketing Science, 18 (4), 504-526. Bell, David R., Ganesh Iyer, and V. Padmanabhan (2002), “Price Competition Under

Stockpiling and Flexible Consumption,” Journal of Marketing Research, 39 (3), 292-303.

Blattberg, Robert C, and K J Wisniewski (1989), “Price-Induced Patterns of Competition,”

Marketing Science, 8 (4), 291-309. Blattberg, Robert C, Richard Briesch, and Edward J. Fox (1995), “How Promotions Work,”

Marketing Science, 14 (3), 122-132.

Page 104: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

97

Block, Stanley B. and Geoffrey A. Hirt (2000), Foundations of Financial Management, Boston: MA, Irwin McGraw-Hill.

Boatwright, Peter, Sanjay Dhar, and Peter Rossi (2004), “The Role of Retail Competition,

Demographics and Account Retail Strategy as Drivers of Promotional Sensitivity,” Quantitative Marketing & Economics, 2 (2), 169-190.

Bolton, Ruth N. (1989), “The Relationship Between Market Characteristics and Promotional

Price Elasticities,” Marketing Science, 8 (2), 153-169. Box, G. E. P. and Wilson, K.B. (1951) “On the Experimental Attainment of Optimum

Conditions (with discussion),” Journal of the Royal Statistical Society Series B, 13 (1), 1-45.

Bucklin, Randolph E. and James M. Lattin (1991), "A Two-State Model of Purchase

Incidence and Brand Choice," Marketing Science, 10 (Winter), 24-39. Bucklin, Randolph E. and Sunil Gupta (1999), “Commercial Use of UPC Scanner Data:

Industry and Academic Perspectives,” Marketing Science, 18 (3), 247-73. Carson, Stephen J., Timothy M. Devinney, Grahame R. Dowling, and George John (1999),

“Understanding Institutional Designs within Marketing Value Systems,” Journal of Marketing, 63(SI), 115-30.

Chandon, Pierre, Brian Wansink, and Gilles Laurent (2000), “A Benefit Congruency

Framework of Sales Promotion Effectiveness,” Journal of Marketing, 64 (4), 65-81. Chandon, Pierre and Brian Wansink (2002), “When Are Stockpiled Products Consumed

Faster? A Convenience–Salience Framework of Postpurchase Consumption Incidence and Quantity,” Journal of Marketing Research, 39 (August), 321–35.

Chakravarti, Amitav and Chris Janiszewski (2004), “The Influence of Generic Advertising on

Brand Preferences,” Journal of Consumer Research, 30 (4), 487-502. Chen, Zhihong, Dan S. Dhaliwal, and Hong Zie (2006), “Regulation Fair Disclosure and the

Cost of Equity Capital,” Social Science Research Network, (September 15, 2006). Available at SSRN: http://ssrn.com/abstract=930724.

Chintagunta, Pradeep K. (1993), “Investigating Purchase Incidence, Brand Choice and

Purchase Quantity Decisions of Households,” Marketing Science, 12 (Spring), 184-208.

Chintagunta, Pradeep, and Sudeep Haldar (1998), “Investigating Purchase Timing Behavior in

Two Related Product Categories,” Journal of Marketing Research, 35 (1), 43-53.

Page 105: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

98

Danaher, Peter J. (1997), “Using Conjoint Analysis to Determine the Relative Importance of Service Attributes Measured in Customer Satisfaction Surveys,” Journal of Retailing, 73 (2), 235-260.

Day, George (1992), “Marketing’s Contribution to the Strategy Dialog,” Journal of the

Academy of Marketing Science, 20 (Fall), 323-30. Day, George, and Liam Fahey (1988), “Valuing Market Strategies,” Journal of Marketing, 52

(July), 45-57 Day, George S. and David B. Montgomery (1999), “Charting New Directions for Marketing,”

Journal of Marketing, 63 (SI), 3-13. Dekimpe, Marnik G. and Dominique M. Hanssens (1995), “The Persistence of Marketing

Effects on Sales,” Marketing Science, 14 (1), 1-21. Dekimpe, Marnik G., Dominique M. Hanssens, Vincent R. Nijs, and Jan-Benedict E. M.

Steenkamp (2005), “Measuring Short- and Long-run Promotional Effectiveness on Scanner Data Using Persistence Modeling,” Applied Stochastic Models in Business and Industry, 21, 409-416.

Dellaert, Benedict G. C., Theo A. Arentze, Michel Bierlaire, Aloys W. J. Borgers, Harry J. P.

Timmermans (1998), “Investigating Consumers' Tendency to Combine Multiple Shopping Purposes and Destinations,” Journal of Marketing Research, 35 (2), 177-188.

Dodson, Joe A., Alice M. Tybout, and Brian Sternthal (1978), “Impact of Deals and Deal

Retraction on Brand Switching,” Journal of Marketing Research, 15 (1), 72-81. Duncan, Delbert, Stanley Hollander, and Ronald Savitt (1983), Modern Retailing

Management, Homewood, IL: Richard D. Irwin, Inc. Dunne, Patrick M. and Robert F. Lusch (2005), Retailing, 5e, Mason, OH: Thompson South-

Western. Fishman, C. (2006), The Wal-Mart Effect: How the World’s Most Powerful Company Really

Works-And How it’s Transforming the American Economy. New York: Penguin. Garrison, Ray and Eric Noreen (2003), Managerial Accounting, 10e, Boston, MA: Irwin

McGraw-Hill. Gedenk, Karen, Scott A. Neslin, and Kusum L. Ailawadi (2006), “Sales Promotion,” in

Retailing in the 21st Century, Manfred Krafft and Murali K. Mantrala, eds. Berlin, Germany: Springer, 345-359.

Page 106: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

99

Goodman, David A. and Kavin <sic> W. Moody (1970), “Determining Optimum Price Promotion Quantities,” Journal of Marketing, 34 (October), 31-39.

Grewal, Dhruv, Howard Marmorstein, and Arun Sharma (1996), “Communicating Price

Information through Semantic Cues: The Moderating Effects of Situation and Discount Size,” Journal of Consumer Research, 23 (September), 148-155.

Guadagni, Peter M. and John D. C. Little (1983), “A Logit Model of Band Choice Calibrated

on Scanner Data,” Marketing Science, 2 (Summer), 203-238. Gupta, Sunil (1988), “Impact of Sales Promotion on When, What, and How Much to Buy,”

Journal of Marketing Research, 25 (November), 342–55. Gupta, Sunil (1991), “Stochastic Models of Inter-Purchase Time with Time Dependent

Covariates,” Journal of Marketing Research, 28 (February), 1-15. Hansen, Jared M. (2007), “The Evolution of Buyer-Supplier Relationships: An Historical

Approach,” Working Paper. (Under review at Journal of Business & Industrial Marketing).

Hansen, Jared M., and Michael J McGinty (2007), “Building Bridges Between Consumption

Research and Practice: The Role of Metaphors,” Working Paper. (Under review at Advances in Consumer Research).

Hansen, Jared M., Sumit Raut, and Sanjeev Swami, (2006), “Retail Shelf Allocation: A

Comparative Simulation Analysis of Heuristic and Meta-Heuristic Approaches,” 8th Triennial Academy of Marketing Science/American Collegiate Retailing Association Retailing Conference, Orlando, FL, 88-93.

Hansen, Jared M., Sumit Raut, and Sanjeev Swami, (forthcoming), “Retail Shelf Allocation:

A Comparative Analysis of Heuristic and Meta-Heuristic Approaches,” Journal of Retailing.

Heller, Laura (2001), “Private Label Plays Off Low-Key, Low-Price Approach,” DSN

Retailing Today, June 4. Hoch, Stephen J., Xavier Dreze, and Mary E. Purk (1994), “EDLP, Hi-Lo, and Margin

Arithmetic,” Journal of Marketing, (58) 4, 16-27. Hunt, Shelby D. (2000), A General Theory of Competition: Resources, Competences,

Productivity, Economic Growth. Thousand Oaks, CA: Sage. Hunt, Shelby D. and Robert M. Morgan (1996), “The Resource-Advantage Theory of

Competition: Dynamics, Path Dependencies, and Evolutionary Dimensions,” Journal of Marketing, 60 (October), 107-114.

Page 107: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

100

Hunt, Shelby D. and Robert M. Morgan (1997), “Resource-Advantage Theory: A Snake

Swallowing its Tail or a General Theory of Competition,” Journal of Marketing, 61 (October), 74-82.

Janakiraman, Narayan, Robert J. Meyer, and Andrea C. Morales (2006), “Spillover Effects:

How Consumers Respond to Unexpected Changes in Price and Quality,” Journal of Consumer Research, 33 (3), 361-69.

Jap, SD (1999), “Pie-Expansion Efforts: Collaboration Processes in Buyer-Supplier

Relationships,” Journal of Marketing Research, 36 (4), 461-75. Kamakura, Wagner A, and Wooseong Kang (2007), “Chain-Wide and Store-Level Analysis

for Cross-Category Management,” Journal of Retailing, 83 (2), 159-70. Kim, Namwoon, Vijay Mahajan, and Rajenda K. Srivastava (1995), “Determining the Going

Value of a Business in an Emerging Information Technology Industry: The Case for Cellular Communications Industry,” Technological Forecasting and societal Change, 49 (July), 257-79.

Kim, Stephen Keysuk and Ping-Hung Hsieh (2003), “Interdependence and Its Consequences

in Distributor-Supplier Relationships: A Distributor Perspective Through Response Surface Approach,” Journal of Marketing Research, 40 (1), 101-112,

Kirmani, Amna and Akshay R. Rao (2000), “No Pain, No Gain: A Critical Review of the Literature on Signaling Unobservable Product Quality,” Journal of Marketing, 64 (April), 66-79.

Kopalle, Praveen K., Carl F. Mela, Lawrence Marsh (1999), “The Dynamic Effect of

Discounting on Sales: Empirical Analysis and Normative Pricing Implications,” Marketing Science, 18 (3), 31-332.

Kuehn, Alfred and Albert Rohloff (1967). "Evaluating Promotions Using a Brand Shifting

Model," in Promotional Decisions Using Mathematical Models, Patrick Robinson, ed. Boston: Allyn and Bacon, Inc., 50-85.

Kumar, V., Vibhas Madan, and Srini S. Srinivasan (2004), “Price Discounts or Coupon

Promotions: Does it matter?” Journal of Business Research, 57 (9), 933-941. Kravitz, Rose Lee (1977), “A New Tool for Measuring GM Performance,” Supermarketing,

32 (6), 8. Leeds, Herbert A. (1976), “Management Has the Information to Improve Return on Inventory

Investment,” Retail Control, 45(3), 34.

Page 108: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

101

Lehmann, Donald R. (2004), Metrics for Making Marketing Matter,” Journal of Marketing, 68 (October), 73–75.

Lodish, (2007), “Another Reason Academics and Practitioners Should Communicate More,”

Journal of Marketing Research, 44 (February), 23-25. Manchanda, Puneet, Asim Ansari, and Sunil Gupta (1999), “The ‘Shopping Basket’: A Model

for Multicategory Purchase Incidence Decisions,” Marketing Science, 18 (2), 95–114. Mason, J. Barry and Morris Mayer (1984), Modern Retailing, Plano, TX: Business

Publications, Inc. McCracken, Grant (1991), Culture and Consumption: New Approaches to the Symbolic

Character of Consumer Goods and Activities. Bloomington, IN: Indiana University Press.

Moriarty, M M (1985), “Retail Promotional Effects on Intra- and Interbrand Sales

Performance,” Journal of Retailing, 61 (3), 27-47. Mulhern, Francis J. and Robert P. Leone (1991), "Implicit Price Bundling of Retail Products:

A Multi-Product Approach to Maximizing Store Profitability," Journal of Marketing, 55 (4), 63-76.

Mulhern, Frank J and Daniel T. Padgett (1995), “The Relationship Between Retail Price

Promotions and Regular Price Purchases,” Journal of Marketing, 59 (4), 83-90. Dr. Miles Medical Co v. John D. Parks and Son Co (1911), 220 U.S. 373. Myers, Matthew B. (2004), “Implications of Pricing Strategy–Venture Strategy Congruence:

An Application Using Optimal Models in an International Context,” Journal of Business Research, Jun2004, Vol. 57 Issue 6, p591

Myers, Raymond H. (1971), Response Surface Methodology, Boston, MA: Allyn and Bacon,

Inc. Myers, Raymond H. and Douglas C. Montgomery (1995), Response Surface Methodology:

Process and Product in Optimization Using Designed Experiments, NY: John Wiley. Nason, Robert W. (2006), “The Macromarketing Mosiac,” Journal of Macromarketing, 26

(2), 219-23. Nelson, Scott A. and Linda G. Schneider Stone (1996), “Consumer Inventory Sensitivity and

the PostPromotion Dip,” Marketing Letters, 7 (January), 77-94.

Page 109: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

102

Nijs, Vincent R., Marnik G. Dekimpe, Jan-Benedict E.M. Steenkamp, and Dominique M. Hanssens (2001), “The Category-Demand Effects of Price Promotions,” Marketing Science, 20 (1), 1-22.

Putler, Daniel S. (1992), “Incorporating Reference Price Effects into a Theory of Consumer

Choice,” Marketing Science, 11 (3), 287-309. Raju, Jagmohan S. (1992), “The Effect of Price Promotions on Variability in Product

Category Sales,” Marketing Science, 11 (3), 207-220. Rappaport, Alfred, (1986), Creating Shareholder Value, New York: The Free Press. Rajan, Madhav V., Stefan Reichelstein, Mark T. Soliman (forthcoming), “Conservatism,

Growth, and Return on Investment,” Review of Accounting Studies. Risch, Ernest H. (1987), Retail Merchandising: Concepts, Dynamics, and Applications,

Columbus, OH: Merrill Publishing Company. Rokkan, Aksel I., Jan B. Heide, and Kenneth H. Wathne (2003), “Specific Investments in

Marketing Relationships: Expropriation and Bonding Effects,” Journal of Marketing Research, 40 (2), 210-24.

Rust, Roland T., Tim Ambler, Gregory S. Carpenter, V. Kumar, and Rajendra K. Srivastava

(2004), “Measuring Marketing Productivity: Current Knowledge and Future Directions,” Journal of Marketing, 68 (October), 76–89.

Shankar, V and L Krishnamurthi (1996), “Relating Price Sensitivity to Retailer Promotional Variables

and Pricing Policy: An Empirical Analysis,” Journal of Retailing, 72 (3), 249-272. Shocker, Allan D., Barry L. Bayus, and Namwoon Kim (2004), “Product Complements and

Substitutes in the Real World: The Relevance of “Other Products’,” Journal of Marketing, 68 (January), 28-40.

Srinivasan, Shuba, Koen Pauwels, Dominique M. Hanssens, and Marnik G. Dekimpe (2004),

“Do Promotions Benefit Manufacturers, Retailers, or Both?” Journal of Marketing Research, 50 (5), 617-29.

Sriram, S., Subramanian Balachander, and Manohar U. Kalwani (2007), “Monitoring the

Dynamics of Brand Equity Using Store-Level Data,” Journal of Marketing, 71 (April), 61-78.

Srivastava Rajendra K., Tasadduq A. Shervani, and Liam Fahey (1998), “Market-Based

Assets and Shareholder Value: A Framework for Analysis,” Journal of Marketing, 62 (1), 2-18.

Page 110: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

103

Sweeney, Daniel J. (1973), "Improving the Profitability of Retail Merchandising Decisions," Journal of Marketing, 37 (January), 60-68.

Swinnerton, Kenneth A. (1997), “An Essay on Economic Efficiency and Core Labour <sic>

Standards,” The World Economy, 20 (1), 73-86. Tolle, Emerson (1976), “Using Management Information Systems to Improve Inventory

ROI,” Retail Control, 45 (3), 2. Useem, J. (2003), “One nation under Wal-Mart”, Fortune, Vol. 147 No. 4, pp. 65. Walters, Rockney G. and Scott B. MacKenzie (1988), “A Structural Equations Analysis of the

Impact of Price Promotions on Store Performance,” Journal of Marketing Research, 25 (1), 51-63.

Walters, Rockney G. (1991), “Assessing the Impact of Retail Price Promotions on Product

Substitution, Complementary Purchase, and Interstore Sales Displacement,” Journal of Marketing, 55 (April), 16-28.

Walton, Sam and John Huey (1992), Made in America: My Story, New York: Doubleday. Warrington, Rosemary (1982), “Measuring Inventory Return on Investment,” Retail Control,

50 (5), 15-21. Webster, Frederick E. (1981), “Top Management’s Concerns About Marketing: Ussue fro the

1980’s), Journal of Marketing, 45 (July), 9-16. Webster, Frederick E. (1992), “The Changing Role of Marketing in the Corporation,” Journal

of Marketing, 56 (October), 1-17. Weitz, Barton A. and Kevin D. Bradford (1999), “Personal Selling and Sales Management: A

Relationship Marketing Perspective,” Journal of the Academy of Marketing Science, 27 (2), 241-54.

Winer, R. (1986), “A Reference Model Choice for Frequently Purchased Products,” Journal

of Consumer Research, 13 (September), 250-256. Woodside, A G and G L Waddle (1975), “Sales Effects of In-Store Advertising,” Journal of

Advertising Research, 15 (3), 29-33. Van Heerde, Harald J., Sachin Gupta, and Dick R. Wittink (2003), “Is 75% of the Sales

Promotion Bump Due to Brand Switching? No, Only 33% Is,” Journal of Marketing Research, 40 (4), 481-91.

Page 111: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

104

Van Heerde, Harald J., Peter S. H. Leeflang, and Dick R. Wittink, (2004), “Decomposing the Sales Promotion Bump with Store Data,” Marketing Science, 23,(3), 317–334.

Vanhuele, Marc and Xavier Dreze (2002), “Measuring the Price Knowledge Shoppers Bring

to the Store,” Journal of Marketing, 66 (October), 72-85.

Page 112: Copyright 2008, Jared Michael Hansen

Texas Tech University, Jared M. Hansen, December 2007

105

Appendix 1: Profitability and Inventory Turnover in GMROII and GMROSI

Turnover 3 Payable 60 days Cost $ 2.50 Sales $ 4.25

Internal

Cost IRR alt

use Shareholder

Cost Retail Sales

Gross Margin

Adj GM

Inventory Turnover GMROII GMROSI

T (0,1) 0 0.17 T (1,2) 0 0.17 T (2,3) 1 0 T (3,4) 1 0 1 2 Average 0.5 0.085 41% 415% 2.10 1.47 8.719941 Turnover 4 Payable 60 days Cost $ 2.50 Sales $ 4.00

Internal

Cost IRR alt

use Shareholder

Cost Retail Sales

Gross Margin

Adj GM

Inventory Turnover GMROII GMROSI

T (0,1) 0 0.17 T (1,2) 0 0.17 T (2,3) 1 1 2 T (3,4)

0.33 0.17 38% 396% 2.40 1.44 9.501 Turnover 6 Payable 60 days Cost $ 2.50 Sales $ 3.75

Internal

Cost IRR alt

use Shareholder

Cost Retail Sales

Gross Margin

Adj GM

Inventory Turnover GMROII GMROSI

T (0,1) 0 0.17 T (1,2) 0 0.17 1 2 T (2,3) T (3,4) 0 0.17 33% 377% 3.00 1.5 11.318 Turnover 6 Payable 60 days Cost $ 2.50 Sales $ 3.00

Internal

Cost IRR alt

use Shareholder

Cost Retail Sales

Gross Margin

Adj GM

Inventory Turnover GMROII GMROSI

T (0,1) 0 0.17 T (1,2) 0 0.17 1 2 T (2,3) T (3,4) 0 0.17 17% 303% 1.20 0.24 3.634

Read: The table here shows four of the many potential points on the response surface that will be constructed. While the first three scenarios show that price promotions can, indeed, potentially provide incremental returns to the retailer and its shareholders (due to inventory turnover acceleration), the fourth scenario show how GMROII and GMROSI differ. In this type of scenario, GMROII indicates that less than one dollar is being generated for each dollar invested in inventory. However, as shown by GMROSI, this type of scenario can still provide a positive return to the investment of the shareholders of the retailer.