The P.R.I.S.M. Project: Measuring In-Store Reach

The audience for in-store marketing can now be measured.

A consortium of manufacturers and retailers spearheaded by the In-Store Marketing Institute has created a tool for measuring in-store consumer reach at the category level. The model was verified by research conducted at four leading retailers in spring 2006.

The project was unveiled September 27 at the In-Store Marketing Expo in Chicago. The announcement is available as a Flash slide presentation (with audio) or Google Video formats.

The measurement model predicts consumer reach by category or area of the store, by retail format, and by day of the week, delivering unprecedented insight into the store as a marketing channel.

The model could prove to be a watershed event for the marketing world, because it allows the store to be compared alongside television, radio and other forms of mass media for its ability to deliver consumer reach. In effect, it could do for the store what the measurement of gross ratings points (GRP) did for television.

While in-store marketing has long been an element of the consumer marketing mix, its potential value as a brand-building vehicle has never been objectively assessed, largely because there hasn't been a way to accurately measure the total reach of a campaign conducted at retail. By establishing a common metric that can be understood by both brand marketers and retailers, this model can eliminate that obstacle.

"This project will unleash new potential for retailers and marketers to create a simpler, more enjoyable shopping experience for people," said Jim Stengel, global marketing officer for Procter & Gamble. "It will give marketers a better understanding of the in-store environment that is certain to have a meaningful impact on the consumer. I applaud the industry collaboration that it took to make this new measurement initiative possible."

"This study has tremendous importance for retailers. Informed by the sophisticated data that this new metric will provide, retailers, for the very first time, now can consider new store layouts and product adjacencies to create an in-store experience that is more relevant, and thus, even more satisfying," said Stephen Quinn, senior vice president of marketing for Wal-Mart. "Our organization has scrutinized the methodology of this study. We find its implications for both retailers and consumers are far-reaching. We are eager to move forward with the next phase that will refine and expand the application of the metric."

What It Is
Officially christened as Pioneering Research for an In-Store Metric (P.R.I.S.M.), the measurement model is the brainchild of the In-Store Metrics Consortium, an ad hoc coalition formed last fall under the Institute's guidance involving retailers Albertsons, Kroger, Walgreens and Wal-Mart and manufacturers 3M, Walt Disney, Coca-Cola, Kellogg, Miller Brewing and P&G.

The consortium's research team began working last spring on the theory that, by predicting in-store traffic, then determining what marketing communications are in the store -- i.e., measuring compliance -- a calculation could be made on the "opportunities to see" a specific communication. This calculation would be similar in scope to the gross ratings points (or GRP) used to estimate the potential audience for TV advertising.

The equation is:

Traffic x Compliance = Opportunity to See

By using existing statistical models that factor out duplicate impressions (accounting for multiple "visits" to the area by the same people) a measurement for consumer reach could be calculated for specific locations in the store.

Traffic x Compliance x Unduplicated Impressions = In-Store Reach

The initial hypothesis was that reliable traffic estimates could be achieved by using data already collected through POS systems and other easily obtainable information, such as store format, product category, units sold and number of category baskets. (The exact formula would be derived on an ongoing basis by directly measuring traffic in a specified set of stores.)

This hypothesis was validated through a pilot study that counted traffic in 63 product categories in 10 stores using infrared sensors positioned in aisles, perimeter locations and store entrance and exit zones.

Adoption of the P.R.I.S.M. model by the industry would deliver a common language for retailers, manufacturers and media buyers to assess the value of retail as a marketing channel and compare its effectiveness to other media such as TV, radio and print. It also would give marketers a metric through which to evaluate the store as a vehicle for generating brand awareness and trial, putting the store on a level playing field with other forms of mass media.

"As the consumer increasingly demands more from marketing messages, and assumes more control over how she chooses to receive them, it's clear that we need to not only understand but master the effectiveness of engagement touchpoints. In-store marketing is a critical communications device and deserves a seat at the table," said Laura Desmond, ceo of media planning agency Starcom/Mediavest USA. "Through this research, based on carefully tested methodology and rich with insights, we now have a powerful metric to build smarter, consumer-centric in-store programs for our clients."

For example, a snack foods brand might find that it can reach the same number of consumers over a one-week period by running a certain number of 30-second spots on broadcast TV, print ads in 20 magazines, or aisle displays in three leading supermarket chains. Based on the campaign's objectives, budgetary guidelines and return-on-investment goals, the brand could then determine which media options make the most sense.

Future Steps
Although the pilot involved just 10 stores in four store formats (supermarket, drug, discount and supercenter), the actual traffic counts validated those derived using the P.R.I.S.M. model so well that researchers are convinced of its reliability. (See attached chart.) A second, more in-depth lead market study is necessary to make the traffic estimates more precise, and more importantly to encompass a broader range of product categories and store formats.

The consortium hopes to begin the next study early in 2007. Members convened Sept. 27 at the In-Store Marketing Expo to unveil the P.R.I.S.M. model and put out a call for participation from additional retailers and manufacturers.

Once the lead market study is completed, the model verified and an effective method of measuring compliance identified, a third-party supplier could quickly move to provide a syndicated service across the industry. (Note: On Dec. 6, the Consortium named VNU's Nielsen In-Store division as its syndication partner. See Related Articles.)

To help understand the issues involved, the Institute prepared answers to what it believes will be frequently asked questions about the P.R.I.S.M. project and its implications for the future.

Why is there a need for this measurement tool?
Consumers have changed in recent decades. As additional options have become available, media usage patterns have changed dramatically. Thus, the task of building and maintaining brand awareness among consumers has become harder. As the pace of life quickens, consumers have less time to devote to any single activity and are demanding that the information they receive from marketers be more relevant and personally beneficial.

These changes in consumer behavior have led marketers and retailers to increasingly embrace the concept of the retail environment as a vehicle for building brand equity in addition to selling products. But this potential can't be fully evaluated without metrics that allow for comparison with other media options.

While the store is one of the oldest marketing environments, it is the least understood from an effectiveness standpoint. Since consumers spend time in stores, it is important to understand how better to provide what they need, when they need it and in the form that best meets those needs. This model provides important insight that will help achieve those goals.

Why hasn't this kind of data been collected before?
Because retail historically has been viewed primarily as a sales channel, with transaction data the standard metric of evaluation. But as noted above, the shopper today is a very different person than she was a few decades ago, and measuring the importance of the store as a marketing channel has taken on new importance.

How do you calculate in-store reach?
Total traffic multiplied by compliance rates produces what the consortium has labeled the "opportunity to see": the number of shoppers who pass by a specific location over a given time period. A formula using known shopper behavior that factors out multiple impressions from the same individuals can then be applied to achieve reach projections.

In relation to overall store traffic, how varied is category-specific traffic?
The level of traffic associated with different product categories varied significantly in the pilot.

How much does category traffic differ from category transactions?
In the pilot, category traffic differed significantly from category transactions both by category and by retailer. Determining that relationship is essential in estimating traffic at a national level.

Can the P.R.I.S.M. model determine the effectiveness of in-store execution?
No, but store traffic data is an integral component in the process. By analyzing audience reach in concert with sales data, marketers will be able to identify the relative effectiveness of various in-store communications.

Can the model be applied outside the food-drug-mass marketplace?
The consortium is confident that it can be applied to all traditional retail channels

Can the model be applied outside the U.S.?
The consortium is confident that the model will apply globally. But it must first complete the U.S. stage before looking to validate the model internationally.

After the lead market study is completed, what is necessary to create a syndicated measurement system?
A third-party supplier will be engaged to manage the lead market study, which will be much more extensive than the pilot study. It would make sense for this supplier to leverage the experience by providing the syndicated service, thus making the transition seamless. As part of the service, the supplier will need to negotiate with retailers for the collection of the data required to calculate in-store reach.

Could any retailer be measured? How often would data need to be collected and in how many stores?
The model has worked across all retail formats that have been tested, and should be applicable elsewhere. Theoretically, any retailer willing to provide data could be measured.

The pilot study evaluated the model in time periods ranging from hourly to weekly. One purpose of the lead market study is to determine the optimal time frame for measurement, as well as the number of stores that must be measured to achieve a reliable sample.

Who would measure compliance? How would they do so?
Various techniques will be examined and the syndicated supplier will play a key role in helping develop an efficient, effective system.

How will retailers benefit from a nationally syndicated system of audience measurement?
A deep understanding of traffic patterns in their stores will give retailers a wealth of opportunities to create more pleasant, informative shopping experiences for customers. That ultimately could lead to increased traffic and sales.

How would manufacturers benefit from a nationally syndicated system of audience measurement?
A deep understanding of traffic levels in stores will allow product manufacturers to assess the value of in-store marketing more effectively, including its value as a vehicle for building brand awareness and trial. It also gives manufacturers and retailers a common language with which to assess marketing opportunities and work together to better meet shopper needs.

Other than for audience measurement purposes, are there other uses for the traffic data that would benefit retailers and/or manufacturers?
There are numerous applications for the data. Retailers will be able to assess areas of high and low traffic density, which could lead to more efficient and effective store layouts, category adjacencies and product selection, to name a few possibilities. The data could also be a catalyst for determining the effectiveness of various methods of in-store communication, and therefore ultimately could lead to improved sales.

Will this model lead to incremental revenue for retailers in the form of "media buys"?
Possibly. The system will establish a common language for retailers and potential advertisers to evaluate opportunities to reach consumers within the store. Retailers will determine what advertising vehicles and communications best meet their objectives, and potential advertisers will be able to evaluate those opportunities accurately.

Will this model lead to increased marketing costs for manufacturers?
The metric should improve collaboration between retailers and product manufacturers by providing a common language for evaluating in-store opportunities to reach consumers. Using that information, they can make informed decisions that best meet their business needs. That potentially could lead to increased allocations in some areas, but also reduced spending in others.

Do you think this will lead product manufacturers to spend less money on TV (and other media) and more on in-store marketing? Will they spend less money on trade promotion?
Brand marketers will continue to evaluate various media and marketing opportunities based on their relative merits for achieving objectives. This metric will put in-store opportunities on a level playing field with other media and allow marketers to better assess the value of marketing at retail.

Will this model devalue traditional in-store tactics (such as displays and signs) in favor of "media" such as in-store TV or radio? Is there a danger that focusing on "advertising to an audience" might distract from "marketing to a shopper"?
The model will measure the number of consumers who have an "opportunity to see" marketing communications in a given area of the store. It will be up to individual retailers and brand marketers to evaluate which specific strategies and tactics are most effective at improving the shopping experience and achieving objectives.

How significant is this development for the future of marketing? Are there other milestones to which it can be compared?
Members of the consortium believe that the ability to measure the in-store audience could be a marketing milestone as significant as the rise of television as an advertising vehicle in the 1950s.

Published: September 2006

Source: In-Store Metrics Consortium