Article Review

DOI: 10.2501/JAR-52-3-339-345 September 2012 JOURNAL OF ADVERTISING RESEARCH 339

INTRODUCTION The current research program quantifies the syn- ergy of the three strongest drivers of positive brand-purchase change:

pricing, in-store display, and television advertising.

The study focused on heavily advertised consumer packaged goods (CPG) brands that averaged upward of $20 million in television advertising within the categories of toothpaste, yogurt, and cereal.

The report summarizes the findings of the first six case studies across six brands with varied mixes of television, price promotion, featured items in a retail circular, and in-store display. It provides

marketers with examples of findings from this single-source, household-level methodology.

Two critical findings:

Within the media industry, it is commonly believed that television and a temporary price- reduction (TPR) promotion should not be used at the same time because television reduces the impact of price reduction. The study has found that to be an inaccurate assumption. For the six case studies, approximately 50 percent of all households exposed to television were also affected by the TPR program and exhibited higher sales increase than the other half reached only by television. Simultaneously using all three marketing tac- tics—pricing, in-store display, and television advertising—maximizes the positive impact

Exploding the Legend of

TV Advertising and Price Promotions The Proper Mix of Price, In-Store, and TV

for Maximum Short- and Long-Term ROI

BILL HARVEY TRA, Inc. [email protected]

TERESE HERBIG TRA, Inc. [email protected]

MATTHEW KEYLOCK dunnhumbyUSA Matthew.Keylock@

us.dunnhumby.com

RITESH AGGARWAL dunnhumbyUSA Ritesh.Aggarwal@

us.dunnhumby.com

NINA LERNER dunnhumbyUSA Nina Lerner@

us.dunnhumby.com

The advertising and marketing communities traditionally have understood television

advertising effectiveness and its relationship to in-store marketing tactics through small-

market research or marketing-mix modeling. Although these studies have improved the

quality of advertising-effectiveness research, they have done little to improve the tools

marketers need to translate those insights on a larger scale and optimize their marketing

those tools and give marketers a more granular understanding of brand-purchase behavior

and the impact of multiple marketing levers on in-store brand sales. This paper leverages

the anonymous household-level purchase behavior data from 60 million households

across the United States and the second-by-second measurement of television-viewing

habits from more than 2 million set-top box households, and the current study applies

actual (non-modeled) single-source, household-level data to demonstrate a methodology

for optimizing the mix of television advertising and in-store marketing.

340 JOURNAL OF ADVERTISING RESEARCH September 2012

EXPLODING THE LEGEND OF TV ADVERTISING AND PRICE PROMOTIONS

on brand sales. The combination of all stimuli at the same time averaged more than 11 times the sales effect of televi- sion alone.

RESEARCH ISSUE Despite the significant impact of price, in- store marketing, and television on brand sales, the authors believe there never has been a methodology for optimizing the three tactics together and measuring how they work in tandem.

Advertising effectiveness has long been a challenge for marketers—a challenge that, in recent years, became more complex as shopper marketing became an import- ant player within the overall marketing matrix. In 2008, some of the world’s larg- est retailers, manufacturers, and adver- tisers invested in the well-conceived but ill-fated PRISM study to understand the impact of in-store display on in-store sales, an attempt to provide a clear indication of just how critical marketers saw shopper marketing in driving brand sales. Lauded as the research that would “transform how we think about in-store consumer com- munication and behavior,”1 if the PRISM experiment had been completed, it might have yielded extremely valuable insight on stimuli-mediated shopper behaviors.

Other research companies have over- laid tests at the store level with television

1 Advertising Age, January 23, 2009, “Nielsen Suspends PRISM Data System,” quoting A. G. Lafley, chairman, Procter & Gamble.

to take snapshots of specific television/ promotion mixes for specific brands. Cumulative or predictive insight from these tests that would enable marketers to implement stronger marketing mixes and develop a sense of the overweening creative variables involved, has just not happened. Yet.

On top of these past inefficiencies, tele- vision often is ignored when the impacts of trade promotions are tested. Instead, promotional programs and television tra- ditionally are measured exclusive from each other. The lack of synergy is rein- forced when both marketers’ and agen- cies’ departments in charge of television and those that oversee trade promotion do not coordinate and communicate.

There never has been a conclusive methodology on which marketers could optimize these variables so as to maxi- mize sales—specifically, for 1-year (“short- term”) or 2-year (“long-term”) return on investment ROI. In fact, many marketers assume that promotion has no long-term effect. If the promotion includes creative, however, such as a feature—for instance, the inclusion of a brand in the retailer’s printed material—it may affect the way shoppers think of a brand. And that effect can be as lasting as any other subcon- scious or conscious impression made on the human mind.

The creative aspects of in-store display are obvious. Yet it seems far-fetched to suppose that a brand may make such a

bad impression on shoppers through such displays as to injure the brand’s position ... or make such a favorable impression as to improve the lasting brand’s standing in that shopper’s mind. Stranger things, however, have happened. The minute one intellectually commoditizes promo- tion by not thinking creatively about the medium, the more likely those dollars will be sub-optimized.

The authors submit

brands need a way to optimize the mix of the main three drivers: pricing, in- store, and television; the industry needs a method to gather insights on how the three tactics interact in a way that is predictable and action- able; and marketers need to recognize the creative dimension of promotion dollars.

The current research study may be the first time that single-source data have been used to co-measure television, TPR, and feature and in-store display. The authors believe that its findings are important for a number of reasons:

This study involved approximately 2,500 stores and looked not only at the impact of in-store display (as other stud- ies have done), but at those sales effects in mix with television, price, and feature in the store’s circular. The study allows the three main driv- ers of ROI (according to marketing mix modeling [MMM]) to be seen acting together at the individual household level. This research makes it possible to accu- mulate generalized learnings beneficial to the industry and marketers by apply- ing a methodology to optimize the pri- mary three drivers of ROI. The research process examined the impact of promotional creative in an

September 2012 JOURNAL OF ADVERTISING RESEARCH 341

EXPLODING THE LEGEND OF TV ADVERTISING AND PRICE PROMOTIONS

attempt to enable practitioners to ana- lyze and identify campaign successes they did not know were successes and, with that knowledge in hand, learn how to repeat and improve upon them.

The current paper summarizes the find- ings of the first six case studies for six brands with varied mixes of television, price promotion, feature in the retail circu- lar, and in-store display.

METHODOLOGY TRA, a media marketing and analytics soft- ware company, and marketing-researcher dunnhumbyUSA established a strategic partnership in 2011 to create an opportu- nity to move from the reliance on tradi- tional panel-based demographics to deliver more-granular, more-effective measure- ment and custom analytical capabilities. This study combined leveraged a database of anonymous customer-purchase behav- ior for more than 60 million households across the United States, encompassing 16 different imprints. TRA’s Media TRA- analytics matched the second-by-second measurement of television viewing hab- its from more than 2 million set-top box households with dunn humby’s customer- purchase behavior data, enabling continu- ous, accurate results regarding campaign success and results.

For each household matched to the set- top box data, a picture emerged of when each household had its shoppers in the store and what they bought. The results also disclosed the day-by-day, brand-by- brand effects of in-store pricing, features, and display.

Three product categories were selected to represent the widest range of diversity in the television/price/in-store mix:

ready-to-eat cereal, yogurt, and toothpaste.

Within each category, two leading brands were chosen to represent variations in this same mix. For each brand, depending on its campaign, an 8- to 14-month window was defined for analysis. The research team then was able to study each brand in a stable campaign for a clean “read.”

Between shopping trips, each measured exposure to brand advertising was seen. Cursory eyeballing of the pattern at the household level offered many intriguing insights in terms of changes in buying behavior in apparent relation to specific television exposures, price discounts, fea- tures in store circulars, and in-store dis- plays. Multivariate analytics articulated these associations into statistically sig- nificant correlations, beta coefficients, and predicted sales volumes when the market- ing stimuli were on and when they were off.

The analysis of single-source (SS) data is very much like MMM carried out at the household level instead of the “market- average” level: the noise is greatly reduced, as there is an obvious regres- sion to the mean in MMM not present in SS. The actual dynamics of stimulus/ response within a household can be seen clearly in SS. In MMM, one is doing infer- ential detective work to draw conclusions from coincidences that, in fact, sometimes just may be coincidences.

GENERALIZED RESULTS Television Advertising Does Not Soften the Impact of Price Discounting Within much of the advertising and mar- keting community, an inverse relationship

between pricing and television advertis- ing historically has been accepted. Con- ventional wisdom has been that television and a TPR should not be used at the same time because, it was assumed, television reduced the impact of price reduction. This finding stemmed from the MMM observation that price elasticity is reduced during persuasive communications, estab- lishing perception and gut belief in brand value.

Within six case studies, the current study found that about 50 percent of all television-reached households also were reached by a TPR. As such, the nonalign- ment of television and price reductions is demonstrated not only to be a perva- sive understanding in the media indus- try but a standard to be a mis-conclusion. The research also found TPR to be a sig- nificant driver overall as it averaged an 11.83-percent increased-sales lift over what television had achieved on its own across these first six cases.

The authors believe that it is wrong to conclude that television advertising should not run coincidental with in-store TPRs just because television allegedly softened the effect of discounting by add- ing brand value perception. Television, in fact, may weaken price as a stimulus, but the TPR retains lots of strength to mul- tiply the sales lift television alone would have.

Using All Marketing Tactics in Tandem Maximizes Sales The current study found that television— in combination with each of three other

342 JOURNAL OF ADVERTISING RESEARCH September 2012

EXPLODING THE LEGEND OF TV ADVERTISING AND PRICE PROMOTIONS

readable stimuli combinations—produced an average sales lift higher than television alone. Using TV in combination with a TPR and feature and display promotions maximized sales response. The combina- tion of all at the same time averaged more than 11 times the sales effect of television alone (See Figure 1).

On average, TV by itself produced a +0.5-percent sales lift over this period. This uplift may seem small at first glance, but it measures individual household levels and, when factored across all of the households reached by television, the effect is significant.

Note that there appears to be an impor- tant synergistic effect. Even double- counting television by adding TV+TPR’s +1.8 percent and TV+Feature’s +0.7 per- cent, the result is only +2.5 percent. It is hard to believe that display by itself added the other +3.2 percent. In essence, the total is more than the sum of the parts. Promo- tional activities for these six brands did not allow for the interaction of TV+Display alone to be measured for each brand to validate this. The sensible conclusion is that there was a compounding effect of the various in-store promotional levers in syn- ergy with television.

These findings are in synch with other findings by the authors in a study of a diet soft drink-brand, where television performance was greatly increased by a combination of that medium plus display (and also by television plus price). It also aligns with other findings that feature appeals to a smaller-sized, more price- sensitive audience and provides lower volume uplift than other promotional activities.

The baseline television sales effect upon which the rest of the stimuli build each brand can be seen in Figure 2. As is typi- cal with single source, there is a very wide range of TV sales effectiveness across brands.

INDIVIDUAL BRAND RESULTS Results for individual brand case studies illustrate the type of findings emerging from this new methodology and technol- ogy. Each table includes all combinations that had sufficient sample sizes. Two of the metrics analyzed are described below. Both of these metrics are percentage increases for the stimuli combination over the parameter for television alone:

% Reach increase: How much the stim- uli combination increased coverage of households (i.e., the percent of house- holds contacted by at least one of the stimuli) and

% Volume increase: How much the stimuli combination increased brand volume (e.g., ounces) used in place of dollar volume to separate out from the negative dollar effects of TPR.

Yogurt Brand A suggests synergy between the various marketing elem- ents. Television+TPR produced a healthy +14-percent volume increase. TV+Feature increased the effect of television alone by +9 percent. Further, television com- bined with feature and display increased the television effect by +27 percent. All in-store stimuli had a greater impact on existing brand buyers. Television

0.5%

1.2%

0.2%

0.1%

0.3%

0.1%

1.0%

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Average

Toothpaste G

Toothpaste F

Cereal D

Cereal C

Yogurt B

Yogurt A

B ra

nd

TV Sales Lift per Household per Week (%)

Figure 2 Six-Brand Average Sales Lift Per Household Per Week

5.7%

0.7%

1.8%

0.5%

0 1 2 3 4 5 6

TV + TPR/Feature/Display

TV + Feature

TV + TPR

TV Only

S ti m

ul i

Average Percent Sales Lift

Figure 1 Average Percent Sales Lift

September 2012 JOURNAL OF ADVERTISING RESEARCH 343

EXPLODING THE LEGEND OF TV ADVERTISING AND PRICE PROMOTIONS

advertising exposure had a greater impact in bringing in new buyers to the brand (See Figure 4).

Yogurt Brand B experienced more mod- est increases from TPR and feature on top of TV—only 3 percent and 2 percent, respectively. Feature plus display, how- ever, added a significant bundled effect, bringing the overall sales lift to +27 per- cent. Presumably, the greater part of the increase in households and volume during this event was due to the display, which had the advantage of capturing the con- sumer at the point of decision.

Similar to yogurt Brand A, all in-store stimuli had a greater impact on exist- ing brand buyers, with display having the greatest impact. Television advertis- ing exposure had a much greater impact in bringing in new buyers to the brand as compared to yogurt Brand A (See Figure 5).

RTE cereal Brand C showed a 12-percent increase in volume over TV alone when TPR was added. Feature added to televi- sion had almost the same size effect (10 percent) as TPR without cannibalizing rev- enues as TPR did. Combining feature and display was the strongest option: the bun- dle brought total impact to +93 percent, nearly doubling overall brand volume (See Figure 6).

RTE cereal Brand D did not benefit much (4 percent) from adding TPR on top of television. Feature was used in so

few store/weeks that it did not support a breakout on its own. The full bundle with display, however, more than doubled tel- evision alone, with a +112-percent volume lift. Combining multiple promotions again appeared to be a stimulus of immense power when added to television. Addi- tionally, all in-store stimuli had a greater impact on existing brand buyers for both RTE cereal brands (See Figure 7).

Toothpaste Brand E saw great benefit from adding TPR on top of television, with a solid 26-percent increase. Feature again was not used in enough store/weeks to read on its own. The TV+Feature+Display brought the total impact of all stimuli com- bined up to a substantial +39 percent over television alone. The study also showed

%

27%

9%

14%

20%

6%

12%

0 5 10 15 20 25 30

TV + Feature + Display

TV + Feature

TV + TPR

S ti m

ul i

% Increase in Household Reach

% Volume Increase

Figure 3 Yogurt Brand A Findings

%

27%

2%

3%

25%

2%

4%

0 5 10 15 20 25 30

TV + Feature + Display

TV + Feature

TV + TPR

S ti m

ul i

% Increase in Household Reach

% Volume Increase

Figure 4 Yogurt Brand B Findings

%

93%

10%

12%

83%

11%

12%

0 20 40 60 80 100

TV + Feature + Display

TV + Feature

TV + TPR

S ti m

ul i

% Increase in Household Reach

% Volume Increase

Figure 5 RTE Cereal Brand C Findings

344 JOURNAL OF ADVERTISING RESEARCH September 2012

EXPLODING THE LEGEND OF TV ADVERTISING AND PRICE PROMOTIONS

that all in-store stimuli had a greater impact on existing brand buyers. The impact, however, was lower than other categories (See Figure 8).

Toothpaste brand saw the highest television-produced sales increase of all six brands in the current study. TPR added only 5 percent. Feature actually had a

negative effect on sales when added to television, lowering brand volume among current purchasers by –3 percent and across the entire market by –14 percent.

Toothpaste is a highly promoted cat- egory with considerable brand switch- ing. The negative results likely were due to coincidental heavy promotional activ- ity by a competitive brand. Another fac- tor: Brand E was not in the feature often; unlike other brands that were in the fea- ture more often, any competitive activity could have affected the analysis of the cur- rent study. Display once again performed admirably with a 14-percent lift. The net positive volume lift across the market- place of the stimuli bundle was +12 per- cent, apparently lower than it would have been without feature.

CONCLUSIONS AND FUTURE IMPLICATIONS The authors believe that the current paper demonstrates, for the first time, a method- ology for measuring and predicting the synergist impact of in-store tactics with tel- evision advertising with household-level data. For this reason, they also believe that this study carries great implications for the future of media planning and marketing spend. Although it does not deliver a blue- print, it does offer a course of action for more accurate media buying and a more strategic allocation of marketing funds.

One important conclusion from this study is the role of creative and com- petitive media and their influence on short-term results. As the current study demonstrated, results will not be predict- able based on any simplistic rules or aver- ages. Each marketer would do well to use single source to study each brand over time so as to be able to learn what is work- ing and what is not and to tune the mix of stimuli to take fullest advantage of where creative is working. Further, this type of analysis can give media planners the tools

%

39%

26%

34%

22%

0 10 20 30 40 50

TV + Feature + Display

TV + TPR

S ti m

ul i

% Increase in Household Reach

% Volume Increase

Figure 7 Toothpaste Brand E Findings

%

12%

14%

–14%

5%

35%

17%

–3%

5%

–20 –10 0 10 20 30 40

S ti m

ul i

% Increase in Household Reach

% Volume Increase

TV + TPR

TV + Display

TV + Feature + Display

TV + Feature

Figure 8 Toothpaste Brand Findings

%

112%

4%

97%

6%

0 20 40 60 80 100 120

TV + Feature + Display

TV + TPR

S ti m

ul i

% Increase in Household Reach

% Volume Increase

Figure 6 RTE Cereal Brand D Findings

September 2012 JOURNAL OF ADVERTISING RESEARCH 345

EXPLODING THE LEGEND OF TV ADVERTISING AND PRICE PROMOTIONS

to reconcile a competitive brand’s media mix during the planning process so that allocations are determined with a high- resolution understanding of the dynamics of the category.

The current study includes the first six brand case studies; as this type of house- hold level analysis continues, the follow- ing questions can now be answered:

Under what conditions, and for which product categories, is there positive synergy between television and in-store and between television and price? How do different consumer groups respond to the various television and in-store promotions? Do they work bet- ter for price-sensitive consumers but not others? What is the effect of different promo- tional stimuli on reaching existing brand buyers versus reaching new or lapsed buyers? Does television reach a different buyer than in-store promotions? Under what conditions are either of these combinations negative anti- synergies? In other words, under what

conditions—and for which product categories—is it wasteful to use televi- sion at the same time as either in-store or price? What decisions made one way based on 1-year ROI would be made differently based on 2-year ROI? For example, in optimizing the three variables, would more television be used based on 2-year ROI than on 1-year ROI?

BILL HARVEY, vice chair and CRO, TRA, Inc. has spent more

than 35 years in the area of media research with special

emphasis on new media. As the strategy head of the

American Research Bureau (now Arbitron), he invented

Century Media and New Electronic Media Science,

third-party research companies serving 70+ of the top

100 advertisers and most major cable and satellite

operators, networks, agencies, and other research

companies. He is a former executive of Arbitron,

Interpublic, Grey Advertising, and OpenTV.

TERESE HERBIG, senior vice president, sales and marketing,

TRA, Inc. brings more than 20 years of packaged goods

experience to TRA. Terese has held positions within

SAMI, Nielsen Marketing Research, and Information

Resources. Herbig has held senior positions in Global

Solution Product Management and Marketing;; CPG and

Retail Marketing and Client Service;; and sales force

development. Follow her on Twitter @therbig

MATTHEW KEYLOCK is senior vice president, new business

development and partnerships at dunnhumbyUSA.

Keylock oversees dunnhumby’s capabilities and growth

in media and marketing effectiveness. Keylock has

been immersed in dunnhumby’s business for more

development of the Clubcard program, its immensely

successful loyalty program and its targeted shopper

communications strategy that have helped to drive

Tesco’s growth. On Twitter: @mattkeylock

RITESH AGGARWAL is Director of Custom Insight at

dunnhumbyUSA, responsible for analyzing customer

purchase behavior to deliver unique, data-driven

insights. He earned a Bachelor ofTechnology in

Mechanical Engineering from the Indian Institute of

Technology in Bombay, India, and a Post Graduate

Diploma in Management from the Indian Institute of

Management in Ahmedabad, India.

NINA LERNER is Associate Director of Analysis at

dunnhumbyUSA, responsible for generating

media insights for dunnhumby’s engagement with

key consumer markets’ clients. Prior to joining

Nielsen Company. She earned a Bachelor of Business

Administration in Business Administration with a

Concentration in Marketing from Emory University, and a

Master of Arts in Quantitative Research Methods in the

Social Sciences from Columbia University.

Copyright of Journal of Advertising Research is the property of Warc LTD and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.