tra and dunnhumby cross media case study

50
TRA Confidential | 1 TRA Confidential Copyright 2012 TiVo Research and Analytics, Inc. 1 TRA Confidential TRA ROI Case Studies Measuring the Impact of TV & Digital

Upload: traglobal

Post on 06-Dec-2014

1.102 views

Category:

Business


5 download

DESCRIPTION

 

TRANSCRIPT

Page 1: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 1

TRA Confidential Copyright 2012 TiVo Research and Analytics, Inc. 1

TRA Confidential

TRA ROI Case StudiesMeasuring the Impact of TV &

Digital

Page 2: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 2

Answering the big questions

The TRA/dunnhumby partnership is uniquely positioned to answer these questions

Can TV’s actual IMPACT ON SALESbe quantified?

What is the SALES LIFT IMPACTof TV and digital?

Can local targeted TV DRIVE BETTER RESULTSthan network TV alone?

Page 3: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 3

80%

TiVo, 20%

4.4 Million HouseholdsAchieving the right composition for nationally

representative sample*

*all data weighted and projected to the US TV population

Charter/Fourth Wall

TRA’s Set Top Box Data Sources

Page 4: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 4

Matched with Best-of-Breed Purchase

Data SourcesTRA is the ONLY company that effectively analyzes and optimizes TV advertising using a massive sample of naturally occurring data

CPG BRANDS RX Brands

CRM Brands

Networks Agencies

TRA Customers

Page 5: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 5

TRA Cross Media Solutions

DigitalBuying

TTI2

MeasuringTRA Cross Media

Measurement

Use TRA TV and purchaser indices to plan/buy digital media:

• Reinforce messaging

• Reach digitally those missed on TV

Purchaser indices used to weight cookies that can be used by any ad serving network to reach the right online audience

Privacy: No household level matching

of TV tuning or purchase data

Using the Right Data for Digital Planning/Buying

Measure the results of your TV and digital

campaigns

Use results to inform/optimize the plan

Single-source cross media solution for:• Measuring the sales impact

of TV and Digital• Understanding attribution to

maximize ROI

Over 100,000 Households Matched

• Auto 100K

• CPG 70K

Privacy: No surfing or PII data collected. Consumer can opt-out using ad choices

Page 6: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 6

Case Studies: TRA is Market Tested & Proven to be Actionable Greater CPG sales from smarter TV allocations led to 4% sales lift across

portfolio, including 9% lift for P&G

Increased ROI by 25% and 35% for two product categories, with data-driven planning for Mars

Measured a 31% ROI and 126% sales lift from a product placement in a popular TV program, compared to exposure to TV ads only, for Garnier Fructis

Proved the impact of Sunovion’s TV advertising and helped to plan for a media buy that maximized exposure at 30% less the cost

Understanding what programs are of greater value to specific auto advertisers by clustering auto owners around program tune-in to help A&E find better selling opportunities

Gauging the impact of TV and digital on sales to help a CPG brand understand that TV drives new customer purchase and digital drives repeat purchase

Matching customer lists to TV Tuning to help a major online retailer understand which advertising creative had the largest effect on sales lift

Page 7: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 7

Introducing dunnhumby

we help companies put thecustomer at the center of every decision.

all our work starts with data

we personalize the experience in communications and the retail environment

we work with clients to change the organization to put the customer at the center through new strategies, tactics, and KPIs

customers repay that loyalty by buying one more product, one more time; this increases sales and profit margin and grows measurablevalue for our clients

Page 9: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 9

ROI Cases Studies: The power of TRA & dunnhumby combined

TV &OnlineExposure

22MM Households

40MM Households

ShoppingBehavior

What is seen, matched with what is purchased.

4 MM Households

Page 10: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 10

Historic Food Product ROI Study: One of the largest ever single source TV sales lift studies

on ~735,000 Households

3.4MM households matched between Comcast subscribers and Kroger shoppers

2.1MM Comcast households in the targeted ad zones

735K Comcast HH in the Kroger continuous panel A

naly

sis

Sam

ple

Page 11: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 11

Campaign MethodologyPrecision purchased-based zone targeting with in-store results

88 zones with high number of category purchasing households and penetration were selected, with a reach of 5 MM households

Category includes food brand and its competitors

Based on a direct household level match between Comcast subscribers and Kroger shoppers.

Selected local zones and networks to target based on historical purchases

Comcast served the TV and digital ads in the zones and networks selected for the campaign

TRA/ dunnhumby measures impact back to point-of-sale(test vs. control / unexposed)

TARGETING HISTORICAL PURCHASES

Page 12: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 12

Key Findings: Targeting works

Sales lift is highest among households exposed to both TV and digital ads

Zone targeting, coupled with purchaser category index, is the key driver for sales lift

The campaign produced a 10% uplift in sales, which is on the high-end of prior studies

TV brings in new customers - 67% of the purchasing household uplift came from new customers (new to brand, new to category)

Digital secures more sales from existing customers

The sweet spot of advertising exposure is 7-10 frequency - sales lift plateaus thereafter

Page 13: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 13

Media Delivery (Analysis Population)

On average, each household was exposed to 12.8 impressions (TV + Digital) over the duration of the

campaign. This included 8.4 TV impressions and 16.4 digital impressions per exposed household.

435k Unique Households (TV + Digital)

‘000s

Thanksgiving break in

campaign

379k Unique Households (TV)

146K Unique Households (Digital)

‘000s

5.6 MM Impressions (TV + Digital)

Thanksgiving break in

campaign

3.2 MM TV Impressions

2.4 MM Digital Impressions

Impressions Unique Households

Page 14: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 14

Advertising Exposure Media Composition

Exposure Group # Households# TV

Impressions# Digital

Impressions

Not exposed to any ads 299 K

Exposed to TV ads only 290 K 2,419 K

Exposed to Digital ads only 57 K 920 K

Exposed to both TV and Digital ads

89 K 745 K 1,474 K

TV ads

only,

67%

Digi-tal ads

only,

13%

TV and Digi-tal Ads,

20%

Analysis Population (Exposed to Ads)

Page 15: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 15

Digital Supplements TV

TV + Digi-tal ,

38%

Light TV,

23%

Digi-tal Only,

39%

Exposed to DIGITAL ADS Nearly two-thirds of those reached by the digital ads had little or

no exposure to the TV campaign

Those lightly exposed to the TV campaign (1-3 impressions), represents only 2% of total TV campaign impressions but 24% of the total digital campaign impressions

Page 16: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 16

Campaign Results: Measurement Methodology

Sales impact is measured during campaign and post-campaign measurement period, for exposed vs. control/unexposed HHs

Control group (matched, not exposed) is selected to “mirror” the exposed group on pre-campaign period sales, volume purchased, trips, demographics, and geography

Sales, penetration, and volume in the campaign and post-period are then compared for exposed and control/unexposed groups

The times periods are dynamically chosen for each exposed HH based on their first exposure, last exposure and a four-week post-campaign period following the final exposure

Continuous Panel: HHs used in the analysis are part of the Kroger Continuous Panel (consistent over time, high level of spend and visits) for the entire pre-campaign, campaign and post-period

Page 17: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 17

Note: Penetration calculated based on buyers / total group (including non-buyers) and is reported at the brand-level

Households exposed to both TV and Digital have the highest sales liftSales Lift – by Media Type

Sales Uplift (All Exposed HHs)

Penetration Sales Uplift (Buying HHs)

0%

2%

4%

6%

8%

10%

12%

14%

1.6%2.6%

10.6%

7.7%

3.5%

12.4%

9.1%

3.8%

Digital Only (54k HHs)

TV Only (274k HHs)

Both TV & Digital (84k HHs)

Page 18: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 18

TV brings in new customersDigital secures more sales from existing customers

Overall TV Only Digital Only Both TV & Digital

0%

20%

40%

60%

80%

100%

38% 33%

48% 50%

31%32%

15%30%

31% 35% 36%

20%

New to CategoryNew to BrandExisting Brand Buyer

100%=

7.4%

7.7%

2.6%

9.1%

Buyer Flow – sources of Penetration Increase

Page 19: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 19

The campaign drove an increase in trial and repeat rates among new brand buyers; leading to higher continuing sales

Series13.0%

3.5%

4.0%

4.5%

5.0%

5.5%

6.0%

4.7%

5.4%

Control Exposed

Trial

Overall 2X > 2X0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

9.1%8.1%

1.0%

9.7%

8.5%

1.2%

Control Exposed

Repeat

Note: Based on 279k HHs that did not purchase Food brand in the 52 week pre-period prior to the campaign

Page 20: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 20

Higher TV ad frequency Drives Sales Lift

Sweet spot is 7-10 frequency, sales lift plateaus above 10 exposuresNote: Sales uplift was not significant for >60 exposures. For “Digital Only” households,

the only significant level was 7-10 exposures, with a sales lift of 9%

Sales Liftby Exposure

Frequency

100% = 273k HHs

15% 22% 21% 16% 11% 9% 3% 1%%HHs =

1 2-3 4-6 7-10 11-15 16-25 26-40 41-600%

4%

8%

12%

16%

4%5%

10%

15%14%

15%16% 15%

TV Only~26% HHs

had a frequency of 11+ TV

ads

Page 21: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 21

Very Price Sensitive (VPS), 77k HHs Consistently purchase products below the

average price point Front Page of ad is very important to this segment Heaviest users of coupons Total price point AND price per volume are

important

Price Sensitive (PS), 176k HHs Consistently purchase products around the

average price point

Least Price Sensitive (LPS), 106k HHs Consistently purchase products above the

average price point Generally less restricted by a budget

Splurge & Save (SS), 52k HHs Behave differently in different parts of the store;

for example, a household might “splurge” in Produce and “save” in Soft Drinks

Households look “Price Sensitive” on average

Most Price Sensitive

Least Price Sensitive

Food brand is a special treat worth buying for “Splurge & Save” shoppers

10%

16%

11%

9%

Sales Lift

Page 22: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 22

Convenience at Home (37k HHs)

Watching the Waistline (40k HHs)

Finest (101k HHs)

Grab & Go Shoppers (26k HHs)

Shoppers on a Budget (32k HHs)

Family Focused (60k HHs)

Traditional Homes (113k HHs)

Sales lift highest Among Segments:“Grab & Go”, “Watching the Waistline”, and “Shoppers on a Budget”

19%

10%

12%

18%

4%

10%

7%SALES LIFTby Segment

Page 23: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 23

Zone sales lift: Category index is the key driver of sales increase

Light BDI (<95)

Medium BDI (95-105)

Heavy BDI (>105)

0%

2%

4%

6%

8%

10%

12%

9%

12%

10%

124k HHs 123k HHs 164k HHs

Light CDI (<=95.5)

Medium CDI (95.5-107)

Heavy CDI (>107)

0%

2%

4%

6%

8%

10%

12%

6%

10%11%

101k HHs 170k HHs 140k HHsNote: CDI = average of (Category Penetration Index , Category Spend Index)

Category penetration index = % HHs in the ZONE that purchased the Category / % HHs overall that purchased the CategoryCategory spend index = Category spend per HH in the ZONE / Category spend per HH overallOverall = average penetration or spend per HH across ALL zonesBDI was calculated in a similar manner to CDI above

Food Brand Lift – by Zone ClassZones are divided into low/ medium/high BDI and CDI

Page 24: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 24

Key Findings: Targeting works

Sales lift is highest among households exposed to both TV and digital ads

Zone targeting, coupled with purchaser category index, is the key driver for sales lift

The campaign produced a 10% uplift in sales, which is on the high-end of prior studies

TV brings in new customers - 67% of the purchasing household uplift came from new customers (new to brand, new to category)

Digital secures more sales from existing customers

The sweet spot of advertising exposure is 7-10 frequency - sales lift plateaus thereafter

Page 25: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 25

Recommendations to Optimize Future Campaigns

Zone Selection Select zones by setting a minimum category sales threshold. Zones with

medium/high CDI have highest sales lift TV Schedule

Reformulate TV schedule to emphasize 7-10 frequency level reach (more networks with less weight per network or more diverse networks)

Network Selection Select networks by utilizing purchaser targets - set a minimum

competitive set rating index. Shopper Segments

Use shopper segments (“Splurge & Save”, “Grab & Go”, “Watching the Waistline”, “Shoppers on a Budget”) as an additional filter for zone and network selection

Media Type Seek strategies to expand digital reach as the combination of television

and Digital generates the highest sales lift

Page 26: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 26

Answering The Big Questions

Yes! A spot cable heavy-up in high CDI areas makes sense. This Comcast 360 campaign occurred at the same time that a national flight was running. Sales lift was 10% higher with consumers targeted by Comcast 360

Can TV’s actual impact on sales beQUANTIFIED?

What is the sales lift impact of TV AND DIGITAL?

Can local targeted TV drive better results THAN NETWORK TV alone?

Yes! This is one of the largest TV & Digital sales lift studies ever conducted, most likely the largest ever in local TV

TV and digital combined, cause the highest sales lift. Each media plays a different role. TV brings in new customers. Digital gets more sales from existing customers

Page 27: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 27

TRA Confidential Copyright 2012 TiVo Research and Analytics, Inc. 27

TRA Confidential

Measuring the In-Store Impact of TV for Brand A and B Wines in Summer

2012TRA |dunnhumby

Page 28: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 28

Measurement universe for each brand made up of around 60K households with continuous shopping behavior and TV viewing data throughout measurement period

717K - households matched between TRA TV panel and Kroger shoppers

150K – limit to households in Kroger continuous panel

62K – limit to households with continuous TV viewing data availability

717K - households matched between TRA TV panel and Kroger shoppers

150K – limit to households in Kroger continuous panel

60K – limit to households with continuous TV viewing data availability

Brand A Brand B

Analysis Sample

Page 29: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 29

Key Findings: TV drives sales

Significant sales uplifts were observed for households exposed to each TV campaign

The results validate the decision to use purchaser targets comprised of brand and competitor brand products to target 2013 advertising Although there are opportunities to refine this in the future to

include particular segments based on price sensitivity and shopper, behavior

Brand A: 43% sales uplift for exposed households (total sales impact of

$1.37MM), driven almost entirely by increase in household penetration

Significant sales uplift was also observed across the entire wine category, particularly Premium wine

Campaign appeared to drive customers into Brand A from outside of the named competitors

Campaign successfully drove trial of the brand, but repeat rate was lower than the control

Page 30: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 30

Key Findings: TV drives sales

Brand B: 38% sales uplift for exposed households (total sales impact of

$395M), driven by a mix of household penetration and $ per household

The campaign drove incremental sales for competitor brands, but not total wine category

The campaign successfully drove trial and repeat of the Brand B brand

Page 31: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 31

While the brand saw the majority of the uplift there was also observable impact across the category, particularly for Premium WineSales Lift for HHs exposed to Brand A ad – by Wine Type

$ Sales Uplift (All Exposed HHs)

Penetration $ per Household0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

43.0%

38.0%

3.2%5.8%

8.6%9.4% 10.7%

1.6%5.0%

6.8%

2.2%4.6%

6.6%

2.1%

Brand A Competitors Premium Wine Bottled Wine Total Wine

>90% for all 66% 72%

Significance

>90% for all 84%

83%

Page 32: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 32

Majority of Brand A customers are active competitor buyers, however the ad was able to bring in customers from outside of this universe

Exposed Control/ Unexposed0%

20%

40%

60%

80%

100%

11% 8%

48%62%

15%8%

17% 15%9% 6%

New to Wine

Existing Bottled Wine Buyer

Existing Premium Wine Buyer

Existing Competitor Buyer

Existing Brand A buyer

Buyer Flow – Post-period Woodbridge purchasers split by pre-period behavior

Note: These results are based on a small sample of Brand A buyers and therefore should only be used directionally.

This validates the decision to target 2013 TV campaign based on purchaser targets of Brand A & competitor brand buyers

Page 33: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 33

The campaign drove an increase in trial rates among new buyers, but not in repeat rates

Series10.0%

0.1%

0.2%

0.3%

0.4%

0.5%

0.6%

0.4%

0.5%

Control Exposed

Trial Rate

Overall 2X > 2X0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

14.0% 13.1%

10.8%

2.3%

10.9%

6.3%

4.6%

Control Exposed

Repeat Rate

Note: Based on 34k HHs that did not purchase Brand A in the 52 week pre-period prior to the campaign

Page 34: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 34

Very Price Sensitive (VPS), 19% HHs Consistently purchase products below the

average price point Front Page of ad is very important to this segment Heaviest users of coupons Total price point AND price per volume are

important

Price Sensitive (PS), 41% HHs Consistently purchase products around the

average price point

Least Price Sensitive (LPS), 27% HHs Consistently purchase products above the

average price point Generally less restricted by a budget

Splurge & Save (SS), 13% HHs Behave differently in different parts of the store;

for example, a household might “splurge” in Produce and “save” in Soft Drinks

Households look “Price Sensitive” on average

Most Price Sensitive

Least Price Sensitive

Brand A creative appears to have resonated with households who are “Price Sensitive” or “Splurge & Save” shoppers in terms of their attitude to price

62%

59%

Sales Lift

Note: Data is only shown where results are statistically significant (>90% significance)

Note: Based on 34k HHs in analysis population exposed to ads

Page 35: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 35

Convenience at Home (11% HHs)

Watching the Waistline (11% HHs)

Finest (16% HHs)

Grab & Go Shoppers (7% HHs)

Shoppers on a Budget (12% HHs)

Family Focused (16% HHs)

Traditional Homes (28% HHs)

In terms of behavior, the creative resonated with households driven by convenience, family or finest products

0.9416721358

0.952363301

2.4539390159SALES

LIFTby Segment

Note: Data is only shown where results are statistically significant (>90% significance)

Note: Based on 34k HHs in analysis population exposed to ads

Page 36: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 36

Brand B: Largest observable sales uplift was for Brand B, however the ad drove overall segment wine volume as well as penetration of competitor brands

Sales Lift for HHs exposed to Brand B ad – by Wine Type

$ Sales Uplift (All Exposed HHs)

Penetration $ per Household0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

38.4%

25.0%

10.7%0.1269112945

0.181255211515.9% 15.5%

3.4%0.4%

6.9%

0.6%

6.6%

Brand B Boxed Wine Competitors Bottled Wine Total Wine

74% 99%

Significance

85%

89% 100% 58%

63%

89% >80%

>90% >90%

>90%

Page 37: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 37

The TV campaign appeared to pull in a larger proportion of bottled wine buyers who were new to boxed wine

Exposed Control/ Unexposed0%

20%

40%

60%

80%

100%

10% 3%

38% 55%

50% 35%

2% 8% New to Wine Cat-egoryExisting Bottled Wine BuyersExisting Competitor BuyersExisting Boxed Wine Buyers

Note: These results are based on a small sample of Brand B buyers and therefore should only be used directionally.

Buyer Flow – Post-period Black Box purchasers split by pre-period behavior

The vast majority of households that try the Brand B brand appear to be existing competitor or bottled wine buyers, validating the purchaser targets selected for the 2013 campaign

Page 38: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 38

The campaign was able to encourage trial of the brand as well as repeat purchase as customers moved into the Brand B wine format

Series10.00%

0.02%

0.04%

0.06%

0.08%

0.10%

0.12%

0.14%

0.10%

0.12%

Control Exposed

Trial Rate

Overall 2X > 2X0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

12.5% 12.5%

0.0%

20.0%18.0%

2.0%

Control Exposed

Repeat Rate

Note: Based on 41k HHs that did not purchase Brand B in the 52 week pre-period prior to the campaign

Page 39: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 39

Brand B Sales uplift was observed across a wide range of exposure frequencies with no clear pattern for optimal exposure

Note: Data is only shown where results are statistically significant (>90% significance)

Brand B Sales Liftby Exposure Frequency

100% = 41k HHs22% 16% 11%%HHs =

2-3 7-10 11-150%

20%

40%

60%

80%

100%

120%

140%

160%

180%

140%154%

117%

Similarly to Brand A, the focus going forward should be on reaching the right customers effectively rather than driving frequency against a wider audience

Page 40: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 40

Very Price Sensitive (VPS), 19% HHs Consistently purchase products below the

average price point Front Page of ad is very important to this segment Heaviest users of coupons Total price point AND price per volume are

important

Price Sensitive (PS), 41% HHs Consistently purchase products around the

average price point

Least Price Sensitive (LPS), 27% HHs Consistently purchase products above the

average price point Generally less restricted by a budget

Splurge & Save (SS), 13% HHs Behave differently in different parts of the store;

for example, a household might “splurge” in Produce and “save” in Soft Drinks

Households look “Price Sensitive” on average

Most Price Sensitive

Least Price Sensitive

The Brand B creative appealed to two polar groups in terms of price sensitivity, reflecting the message of great quality at a big discount over comparable wines when purchased in volume

1.5980328831

2.1076762901

Sales Lift

Note: Data is only shown where results are statistically significant (>85% significance)

Note: Based on 41k HHs in analysis population exposed to ads

Page 41: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 41

Both campaigns were able to drive significant sales uplifts for their respective brands This sets a solid benchmark from which to measure performance of

the 2013 campaigns

The results add weight to the decision to base the 2013 buy on brand and competitor purchaser targets TV is able to bring in new customers to both the brand and category,

however existing brand and competitor purchasers are the strongest group from which to build targeting

This is the third study that proves TV brings in new customers from "outside a brand's immediate competitive waters"

The creative for each campaign appeared to resonate with the intended target groups The challenge going forward is to reach these targets efficiently and

effectively

Key Findings: Targeting works

Page 42: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 42

TRA Confidential Copyright 2012 TiVo Research and Analytics, Inc. 42

TRA Confidential

Q&A|dunnhumby | TRA

Page 43: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 43

Appendix

A. Measurement MethodologyB. Additional Campaign Results

Page 44: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 44

A. Introducing TRA: TRA Matches TV Tuning, CPG Purchase Data, Rx Prescriptions and Auto Registrations

TRA directly matches purchases and the TV tuning of the exact same household (HH) with huge samples. Auto: TRA matches 115 Million HH automotive

registrations to HH TV tuning via 85% Cable/15% TiVo set-top-boxes (4.2MM total)

Household Demographics: Via Experian by HH (115MM)

CPG: TRA matches supermarket purchase by HH (50MM) to HH TV tuning via set-top-box data

Pharma: TRA matches IMS 70 Million HH Rx prescriptions filled to HH TV tuning

CRM: TRA matches tuning data with proprietary purchaser databases from advertisers

Page 45: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 45

C. Measurement Methodology

Time periods are dynamically chosen based on first and last exposure to the campaign as well as the number of weeks post final exposure Pre period: 52 weeks prior to the start of the campaign Campaign period: Weeks the campaign was executed Post period: 4 weeks post the campaign

Nov 12 Dec 16

= Exposure

Final Post period=Campaign Period + Post Period

Jan 13

Page 46: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 46

Food Brand TV Exposure FrequencyMost impressions occur in 4-20 range

# HH # TV Impressions % HH % TV Impressions

Low (1-3 imps) 142 K 263 K 37% 8%

Medium (4-20 imps) 204 K 1,796 K 54% 57%

High (>20 imps) 34 K 1,105 K 9% 35%

Low High

% o

f H

ou

seh

old

s % o

f TV

Im

pre

ssio

ns

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 3941+

0%

2%

4%

6%

8%

10%

12%

14%

16%

0%

2%

4%

6%

8%

10%

12%

14%

16%

HHs TV Impressions

Page 47: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 47

Food Brand Digital Exposure FrequencyMost impressions occur in 5-50 range

# HH # Digital Impressions

% HH % Digital Impressions

Low (1-4 imps) 50 K 110 K 34% 5%

Medium (5-50 imps) 85 K 1,427 K 59% 60%

High (>50 imps) 10 K 857 K 7% 36%

Low High

% o

f H

ou

seh

old

s%

of D

igita

l im

pre

ssio

ns

1 3 5 7 91

11

31

51

71

92

12

32

52

72

93

13

33

53

73

94

14

34

54

74

95

15

35

55

75

96

16

36

56

76

97

1+

0%

2%

4%

6%

8%

10%

12%

14%

0%

5%

10%

15%

20%

25%

30%

HHs Digital Impressions

Page 48: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 48

Brand sales of exposed HHs and their matched control / unexposed HHs match closely in the pre-period, for all exposure segments

Sales in 4 week periods, Analysis Population

$0

$20,000

$40,000

$60,000

$80,000

$100,000

$120,000

Nov-11 Dec-11 Jan-12 Feb-12 Mar-12 Apr-12 May-12 Jun-12 Jul-12 Aug-12 Sep-12 Oct-12 Nov-12 Dec-12

TV Only

Both

Digital Only

During and after campaign

Pre-period

Page 49: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 49

TV + Digital Exposure Frequency – 72% of homes reached in three quadrants (Medium TV/Medium Digital)

Digital Exposures

TV Exposur

es

Low Medium High

Low

11 K HHs (12%)

3% TV imps2% Dig imps

20 K HHs (22%)

5% TV imps23% Dig imps

2 K HHs (3%)1% TV imps

14% Dig imps

Medium

17 K HHs (19%)

20% TV imps2% Dig imps

28 K HHs (31%)

33% TV imps31% Dig imps

3 K HHs (4%)4% TV imps

19% Dig imps

High3 K HHs (3%)13% TV imps0% Dig imps

5 K HHs (5%)20% TV imps5% Dig imps

1 K HHs (1%)2% TV imps3% Dig imps

Page 50: TRA and dunnhumby Cross Media Case Study

TRA Confidential | 50

Convenience at Home - 37k HHs: motivated by foods that can be produced at a moment’s notice and not spoil. Their need for convenience foods can be driven by various factors, like being paid only once a month, hence having fewer large shopping trips

Watching the Waistline - 40k HHs: choose their products based on their fat/calorie content, with the main motivator being weight loss or maintaining a healthy weight. They look for the best quality when making their grocery purchases, focusing on diet/organic foods, fresh fruits and vegetables

Finest - 101k HHs: quality food is central to the “Finest” segment. They do not tend to buy pre-packaged foods’ instead they will spend their money on fresh, quality foods. They shop in the Gourmet, Home Cooking, Organic and Vegetarian sections

Grab & Go Shoppers - 26k HHs: motivated by the speed at which they can get in and out of the grocery store. They make more trips than the avg. HH and spend less per trip. They also tend to buy items in smaller packages.

Shoppers on a Budget - 32k HHs: governed by one factor – price. They are careful not to impulse buy and use shopping lists to save money

Family Focused - 60k HHs: little time to spend shopping. They tend to buy a diverse set of products, as they do not have the time to ‘cherry pick’ offers across stores. They buy in Healthy Snack, Sport, Kid and Baby sections

Traditional Homes - 113k HHs: the driving force is the ability to prepare good quality home cooked meals. A typical meal consists of meat and vegetables. They opt for fresh seafood and meat rather than pre-packaged.

Dunnhumby Shopper SegmentsHH’s in Exposed To The Campaign