MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution...

31

Transcript of MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution...

Page 1: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts
Page 2: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

MARKETING PERFORMANCE MANAGEMENT: ATTRIBUTION

OR MARKETING MIX?

Thought Leadership by

2

Page 3: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

Srishti Gupta President

IRI Media Center of Excellence

Joy Joseph Practice Leader, Marketing Productivity

IRI Strategic Analytics

TODAY’S PRESENTERS

3

Page 4: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

TODAY’S AGENDA

Why is this debate gathering momentum?

The case for and against Attribution

Pros and Cons with MMM

[MMM + Attribution]= Marketing Productivity Excellence

Vision for the future

4

Page 5: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

TODAY’S AGENDA

Why is this debate gathering momentum?

The case for and against Attribution

Pros and Cons with MMM

[MMM + Attribution]= Marketing Productivity Excellence

Vision for the future

5

Page 6: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

INFORMATION EXPLOSION HAS CREATED A DIVERSE AND WELL INFORMED CONSUMER

Health Benefits Occasions

Lifestyle

Democracy of information access has led to a highly evolved and diversified consumer landscape - targeted marketing has

become indispensable for consumer relevancy

Digital revolution has also created complex shopper journeys - imperative to integrate online, offline and touch points to engage across various

stages of the Shopper journey 6

Page 7: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

PROLIFERATION OF MEDIA TOUCH POINTS HAS LED TO A COMPLEX CONSUMER JOURNEY

Digital ad optimization supported a far more complex engagement model, individual level sales conversion rather than aggregate sales contributions were required - enter Attribution Modeling

A parallel measurement environment was initially OK because Digital stakeholders initially operated in a separate environment from traditional media, but not for long…

Speed

Granularity

7

Page 8: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

THESE SEPARATE WORLDS COLLIDED AS ADVERTISERS RAMPED UP ON DIGITAL INVESTMENTS, FREQUENTLY AT THE COST OF TRADITIONAL MEDIA

Digital and traditional media stakeholders began finding themselves in the same conference room with increasing frequency - a comparison and even bake-off between the two approaches was inevitable.

DIGITAL AGENCIES

PUBLISHERS/ AD NETWORKS/ EXCHANGES

CRM/ DIGITAL MEDIA LEAD

INSIGHTS

BRAND MANAGEMENT

TRADITIONAL AGENCIES

MMM Sandbox

Attribution Sandbox

Buy-side Sell-side

8

Page 9: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

TODAY’S AGENDA

Why is this debate gathering momentum?

The case for and against Attribution

Pros and Cons with MMM

[MMM + Attribution]= Marketing Productivity Excellence

Vision for the future

9

Page 10: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

IN AN INCREASINGLY ADDRESSABLE WORLD, ATTRIBUTION EPITOMIZES MICRO-MARKETING

Un-Exposed

EXPOSED

Conversions (“Lift”)

Display + Search

Social + Search

Social + Search

Cro

ss-P

latfo

rm S

yner

gies

Lifestyles Purchase Behavior Demographics

Post-campaign Diagnostics

10

Page 11: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

HYPER-GRANULARITY UNCOVERS ARBITRAGE OPPORTUNITIES WITHIN THE LONG-TAIL OF DIGITAL OPTIONS

Leaderboard

Rectangle S

kysc

rape

r Internalize pay-for-performance: Align growth opportunities with the

optimal execution mix

11

Page 12: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

THIRD-PARTY MATCHING ALLOWS ATTRIBUTION PLATFORMS TO INTEGRATE ACROSS A WIDE RANGE OF DATA SOURCES

HH

12

Page 13: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

NEAR-REAL-TIME MEASUREMENT TO SUPPORT PROGRAMMATIC INITIATIVES

Mid-campaign Attribution Read

Optimize/Course-correct

Post-campaign Diagnostics/ Lessons Learned/ Set-up for next campaign

4-8 Week Cycle-time

DMP/ Conversion Metrics

Programmatic/ DSP/Ad Exchanges

DMP/ Conversion Metrics

13

Page 14: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

ATTRIBUTION COMES WITH ITS OWN ISSUES THOUGH…

• Attribution largely ignores offline touch points, resulting in potentially inflated ROAS estimates

• Big data doesn’t mean “all data”- Integrating micro-data across multiple sources with varying quality and inconsistent matching across multiple third party sources potentially drive biased campaign lift estimates

• Attribution can be too short-term focused, optimizing over weeks rather than quarters, so cannot address longer-term marketing planning decisions

14

Page 15: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

TODAY’S AGENDA

Why is this debate gathering momentum?

The case for and against Attribution

Pros and Cons with MMM

[MMM + Attribution]= Marketing Productivity Excellence

Vision for the future

15

Page 16: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

MMM ENABLES PORTFOLIO STRATEGY BY OPTIMIZING ALLOCATION BY BRAND AS WELL AS BY PLATFORM

15

20

30

35

18

36

18

28

Brand A

Brand B

Brand C

Brand D

Current Optimal

Spe

nd A

lloca

tion

Cro

ss-P

latfo

rm A

lloca

tion

16

Page 17: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

BRAND CROSS-EFFECTS & HALO CAN BE LEVERAGED FOR EFFICIENT CROSS-BRAND CAMPAIGNS

A B C D

A

B

C

D

Halo Impact

Adv

ertis

ed B

rand

Cro

ss-B

rand

Hal

o

Sou

rces

of M

ktg.

Sal

es

17

Page 18: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

MMM ALSO SUPPORTS STRATEGIC PLANNING PROCESSES…

Long

-term

Bra

nd E

quity

Dem

and

Pla

nnin

g

18

Page 19: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

DETERMINE OPTIMAL SPEND LEVELS ACCOUNTING FOR SATURATION AND SIMULATE MULTIPLE STRATEGIES FOR SUPPORTING INNOVATION…

New

Pro

duct

Mar

ketin

g

Rig

ht S

pend

19

Page 20: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

MOST IMPORTANTLY, MMM ALSO SUPPORTS OFFLINE MEDIA PLANNING…

20

Page 21: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

MMM DISADVANTAGES

• Digital can present challenges for MMM, especially where granular data (down to markets and trade areas) is infeasible

• Hyper-targeted and addressable campaigns do not have the scale to register accurately in MMM

• Pure MMM takes a one size fits all approach, but marketing responsiveness varies across consumer groups

• Synergies, especially across digital campaigns are difficult to read without single source data (down to individual audience level), resulting in underestimation of cross-platform campaigns in MMM

• Finally, the broad and encompassing demand measurement MMM undertakes comes at a cost; long cycle times - model updates can often take 8 or more weeks from data gathering to insights/execution

21

Page 22: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

LACK OF INTEGRATION CAN IMPACT INSIGHTS AND STRATEGY IN MULTIPLE WAYS...

Case Study:

12-week Cross-Media Campaign for a OTC Health Brand - MMM indicated the media portion was around 70 percent of the lift suggested by the attribution model (11.6 percent vs. 16.5 percent).

A major in-store shopper initiative at a large regional retail chain was skewing Online Video up in the attribution study, while MMM was underestimating the highly targeted Display campaign.

MTA Online Video Display TV Sales Lift 5.7% 3.6% 7.2%

MMM Online Video Display TV Promo Sales Lift 2.1% 2.9% 6.6% 4.1%

22

Page 23: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

WHAT DO YOU DO WHEN FACED WITH APPLES AND ORANGES?

23

Page 24: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

TODAY’S AGENDA

Why is this debate gathering momentum?

The case for and against Attribution

Pros and Cons with MMM

[MMM + Attribution]= Marketing Productivity Excellence

Vision for the future

24

Page 25: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

DEFINE THE HORIZON & CADENCE…

Both MMM and Attribution Modeling are great for the purpose they are intended for:

MMM guiding overall marketing strategy

Attribution Modeling guiding tactical campaign performance

25

Page 26: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

DESIGNING THE INTEGRATED ANALYTIC MODEL

• Offline and non-marketing drivers’ impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts

• Lifts from Digital and other targeted campaigns measured in Attribution models are used in MMM-driven budget allocation and portfolio optimizations to more accurately represent these

26

Page 27: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

SUCCESS DEPENDS UPON EXECUTION…AND ADOPTION

Page 28: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

TODAY’S AGENDA

Why is this debate gathering momentum?

The case for and against Attribution

Pros and Cons with MMM

[MMM + Attribution]= Marketing Productivity Excellence

Vision for the future

28

Page 29: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts
Page 30: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

SOME PREDICTIONS FOR THE VERY NEAR FUTURE…

Immediate future will see a more cogent integration of MMM and Attribution

A greater focus on targeting will push measurement down to at least household-level, if not individual consumer level

Now that we have stopped being awestruck by the sheer amount of data that is creating “insights”, it is time to address issues like data bias

Focus will shift to predictive accuracy vs. backward-looking campaign scorecard

Advanced probability based models will be explored to compensate for data biases and for understanding non-addressable offline media

Page 31: MARKETING PERFORMANCE - IRI · impact from MMM are leveraged as covariate controls in Attribution Modeling to more accurately quantify Digital/addressable campaign lifts • Lifts

THANK YOU!

[email protected]

[email protected]

QUESTIONS?