Post on 01-Nov-2014
description
Marketing Metrics forDigital Retail MediaExperimental Design for Campaign Measurement
Bio
John MorganFounder and President
• 20-year career in media and marketing research
• Advertising, media, retail and technology expertise
• Former VP, Research & Shopper Insights at InStore Broadcasting
• Designed media planning systems at Nielsen IMS
• Media planner at Hal Riney & Partners
Overview
• Marketers are increasingly demanding accountability– Proof of performance– Consumer insights– Ad optimization– Return on investment
Digital Retail Media:A Perfect Laboratory
• Located at point-of-purchase• Can be interactive• Addressable• Low-cost production• Efficient• Category exclusivity
Experimental Designs for ROI Analysis
• Matched panel design– True experimental design that measures
the effect of an advertising variable compared to a control during two observation periods
• Latin square design– Used to measure the effect of multiple
advertising variables without the use of a control
Matched Panel Experiments
• A method for testing the effect of a campaign variable (advertising creative) on brand metrics (brand equity, purchase intent, sales) in a real-world environment
• Small scale testing minimizes the risk and costs associated with a full roll-out
Matched Panel Methodology
• Locations are matched on traffic, category sales, demographic variables
• Locations are grouped – Control cell (no advertising)– Test cell(s) (advertising variables)
• Sales are compared on a pre vs. post / test vs. control basis
Match!
Control Store
Test Store
Matched Panel Example:Convenience Store TV
• Product: Energy drink brand• Venue: Regional c-store video
network• Objective: Positive ROI• Measure: Net incremental
revenues (increased profit minus media and production costs)
Matched Panel Example:Convenience Store TV
Net
Rev
enue
s
Difference in Sales
Pre-Period Campaign-Period
+ $6,000
Matched Panel Example:Convenience Store TV
• Advertising ROI AnalysisNet Revenue Difference $6,000- Media & Production Cost $5,000= Return on Investment $1,000
Latin Square Experiments
• A method for testing the effect of multiple campaign variable (advertising creative) on consumer response without using a control
• Minimizes the number of venues needed for testing
A
B
C
Ad
Latin Square Methodology
• Locations do not have to be matched, but should be representative
• Each advertising variable is introduced in every test cell during a different period of time
• Customer response is measured for each combination of advertising variable and time
Latin Square Example
• Product: Checking account promotion for a national bank
• Venue: Doctor’s office waiting area video network
• Objectives: QR code scans direct to account application, new checking accounts opened
• Measures: Number of QR code scans
Latin Square Example:Comparing Outcomes
# QR Scans Week 1 Week 2 Week 3
Office I Creative A (15) Creative B (20) Creative C (10)
Office II Creative B (25) Creative C (15) Creative A (20)
Office III Creative C (20) Creative A (25) Creative B (30)
Total QR ScansCreative A: 60Creative B: 75Creative C: 45
Creative B is most effective!
John MorganFounder and President
Retail Media Insightsjohn@retailmediainsights.com
OUR SERVICES Audience Measurement and SegmentationAdvertising Sales and Marketing StrategyConsumer Insights for Content GuidanceAdvertising Campaign ROI Analysis
RESEARCH METHODSSyndicated Research SourcesMatched Panel AnalysesConsumer Intercept SurveysObservational Studies
CLIENTS WE SERVEAdvertisers and AgenciesMedia Network OperatorsRetailers and Venue OwnersInvestors
The Proof is in The Numbers
www.retailmediainsights.com