Best Practices in Shopper Marketing Measurement, Part 6: Directions for the Future

The following is the final installment in a six-part series examining best practices for the measurement of shopper marketing. This article explores directions for the future. Part 6: Directions for the FutureThe ability to compare marketing activity across platforms has been a highly desired but unattainable goal for a generation of marketers. Today, however, the Holy Grail the industry seeks is on the horizon – if not within immediate reach.So what specifically is this Holy Grail? According to the Coalition for Innovative Media Measurement (CIMM), “Single source data for passive media measurement is the Holy Grail for cross-platform advertising effectiveness.” (See "Best Practices in Cross-Platform Advertising Effectiveness Measurement" at right.) This is a fairly technical statement that deserves to be explained in detail. First of all, the term single source refers to tracking a single shopper, recording all of her advertising exposures and all of her purchases, for a defined period of time. The ideal single source data stream would provide a true 360-degree picture of a shopper’s day from a media-exposure standpoint and also would include all of her purchase behavior.So far, the best available models can track a shopper’s exposure in three channels: online, television and in-store. These areas constitute the term cross-platform in CIMM’s statement. While this new approach is a great leap forward, it still excludes numerous other channels of media exposure such as radio, outdoor and mobile.Gathering the data works roughly this way: Purchase behavior is tracked using scanner data derived from credit/debit card transactions. Each shopper is assigned an anonymous tag, which is then matched to online activity connecting that shopper to a specific online surfer. So the single source becomes a specific shopper. In a final step, certain households that have their TV viewing monitored are matched to the anonymous tags, completing a single source, in-store/out-of-store tracking system. (See Venn diagram.)An enormous advantage to this approach is that it can report recorded behavior, which is captured digitally as the passive observations noted in CIMM’s statement. The individual isn’t a “respondent,” so there are no questionnaires to fill out or incentives to dangle. Instead, this technique is one of the first to take advantage of “big data.”Big data is the huge set of “exhaust data” that is emerging. When consumers use the Internet, they leave behind trails of data that serve as a record of their behavior. There are a lot of trails generated by both man and machine. Like consumers, machines also generate exhaust data, which includes monitoring records of every stripe, as well as machine-to-machine communication trails like electronic data transfer messages. Because these data include passively observed records of consumer behavior, they are more accurate than asking respondents to remember and report their activity.While this new methodology is exciting, there are issues with the approach. First, as previously noted, many media platforms are missing. Notably absent at this point are magazines, newspapers, outdoor media and radio. Further, the industry is still working to link TV and retail consumption to digital platforms such as mobile applications. At this stage, the most interesting addition to the current database would be smartphone user behavior (which may not be too far away).Until all these links are established, media consumption is ascertained using old-fashioned surveys to fill in the gaps on a per-person basis.There is another big issue with this nascent approach. While these three databases are robust enough to find and track individuals, sample sizes sometimes inadequately represent populations. Weighting and other statistical techniques can be used to compensate for sample sizes until more observations are collected, but this is a bit of a drawback in the short term.Beyond these two issues, there is the social (some might even say “moral”) issue of privacy. Research companies engaged in tagging and tracking are quick to assert that they are scrupulous in their use of anonymous tags. Additionally, more than one company is always involved in matching the data, so a specific individual’s identity can’t be revealed, even from a logistical standpoint. The tags and data hand-offs are important, therefore, because they are crucial to preserving anonymity of individuals; otherwise, this kind of analysis could be perceived as surveillance.Marketing Mix AnalysisMarketing mix models have been used since the 1980s to understand the impact of various marketing tactics on sales. Until now, marketers haven’t had access to this enormous exhaust data, which lets them observe media consumption in conjunction with purchase behavior on a mass scale. So older mix models used statistical techniques to examine the correlation between sales and various independent marketing variables. This was a forensic exercise in which analysts would infer the causes of sales increases based on advertising impressions, spending and other factors. The “shoppers” were a gross, unified force behind the statistical relationships.The picture changes when marketing mix models get the high-octane input of single source data. Models are still run, but they are used to analyze the relationship between real “opportunity to see” ad exposure and an actual purchase. How likely is a shopper to buy Brand X after exposure to Online Ad Y? After exposure to Online Ad Y and TV Ad Z? These kinds of questions get very accurate, reality-based answers due to the extensive records available today.How We’ll Roll: New OutputsSingle source data will revolutionize how companies go to market and gauge success by delivering five prominent new benefits to shopper marketers.1. An Enhanced Ability to Rationalize and Optimize Investments: Without even tapping into modeling or analytics, it is possible to simply examine tracked shoppers to reveal the percentage of households that were exposed to various media and their actual contribution to revenue in the same time period. This is not modeled behavior but observed, in-market behavior. Additionally, it is possible to model the incremental household penetration generated by each marketing activity.Armed with this information, a marketer can optimize investments across platforms because the technique reveals the relationship between marketing activities and sales response much more accurately than models of the past. This is a thought-provoking advantage, because it inches marketing closer to an era of “perfect information” – not quite grasping the Grail, but nearing it.2. An Accurate Picture of In-Market Responsiveness by Shopper Segment: Because shopper behaviors are observed and not generated through statistics, it is possible to separate groups of shoppers into multicultural, generational or other segments. Marketing responsiveness then can be read and compared.Such in-market, passive observation is likely to be more revealing of real-world responsiveness than other methodologies such as central location testing, a technique that ultimately may be relegated to prelaunch diagnostics and communications refinement.3. An Assessment of Media Opportunity Costs: Interestingly, this technique allows marketers to look at the buying rates among shoppers who were exposed to advertising and those who were not. It also can show the buying rates of those not reached through online or television advertising. For example, category buyers who were not exposed may constitute an opportunity cost (lost sales) that can be pursued in subsequent media plans – when tracked consumption patterns could identify optimal placement. That is really different.4. The Ability to Better Optimize Media Plans: The “rule” of three exposures as the optimal frequency has been accepted for years. Single source data inputs may not completely debunk that, but they can create a more fact-based discussion of optimal frequency, as seen in the sample chart below.While this kind of chart has been available in the past, it was harder to gather the data, so only the biggest brands could produce them. Alternatively, this graph could have been generated through modeling, which theoretically at least makes it less accurate. As cross-platform analysis becomes more prevalent, the data should become more accurate and less expensive, allowing even smaller brands to get better decision-making tools.5. An Understanding of the Role Each Vehicle Plays in Driving Sales: Because opportunity-to-see data is tied to sales figures, it is possible to understand the role each marketing element has had on the business. Now, marketers can look at the real impact of TV or in-store, for instance. These vehicles play different roles that will be tracked – not merely modeled. (See chart below.)This is great news for shopper marketers, who are looking to create effective messages at different points along the path to purchase. Strategy and creative teams can leverage this information to drive specific behaviors at distinct touchpoints.Big Data, Big ImplicationsSince the turn of the millennium, shopper marketers have been calling for their companies to “break down the silos” within entrenched legacy departments, which were developed in response to the marketplace needs of 20 years ago. What was an efficient practice in 1995 is hampering growth in 2012.“We need to break down the silos. These conversations will not be easy because of the mindset that trade is trade, brand dollars are in one bucket and shopper dollars [are] in another,” Jim Fuqua, Supervalu’s then-director of shopper marketing, told Shopper Marketing in early 2011. Booz & Co. has similarly called for the change: “To maintain its growth and fulfill its promise, shopper marketing must evolve beyond a siloed, tactical practice and become a strategic capability that is better integrated with other major investments and across the marketing and media ecosystem.” (Shopper Marketing 3.0: Unleashing the Next Wave of Value. Booz & Co./GMA, 2010)Repeated calls like those have met with a slow – even obstinate – industry response. While intentions are often sincere, reorganizing can be painful and, therefore, is often resisted. But big data will force the change. According to Bill Pink, senior partner at Millward Brown, New York, “The data itself will break the silos. You can’t have data that speaks to everything and people who don’t talk to each other.” (Hear more of Pink's outlook in "The Future of Marketing Measurement" in Related Videos.)Patrick Fitzmaurice, principal of The Capré Group consultancy in Atlanta, envisions a new role for customer-facing talent that “will be very different from the rates-and-dates mentality of the past. The new team leads will manage a joint venture between brands and retailers, going after pockets of demand and allocating resources to maximize returns very specifically.”The effective use of big data promises to answer the need to rationalize marketing investments by tracking individual shoppers as they are exposed to multiple marketing elements. Further, it potentially can erode silo walls because the data will “speak to everything,” forcing heretofore discipline-focused marketers to do the same.It also may level the playing field somewhat between large and small brands – and large and small marketing efforts. If the cost of digital data output declines (as it has consistently in the past), the cost of measuring campaigns may become affordable for most marketers. The implication is that smaller, savvy players may better be able to give big brands a run for their money.While these data techniques are still nascent, one thing is certain: Big data is coming. It’s time for shopper marketers to prepare.About the AuthorLiz Crawford has more than 20 years of brand management and consulting experience with a concentration in strategic innovation. Over the last few years, Crawford has focused on developing integrated shopper marketing strategies for Fortune 500 clients. Currently, Crawford is an analyst and contributing writer for the Path to Purchase Institute. McGraw-Hill released her book, “The Shopper Economy,” in March.JWT/OgilvyAction Inc., conducting business under the OgilvyAction and JWT Action brands, is a fully integrated, end-to-end shopper marketing and experiential marketing agency with main offices in New York, Chicago and Akron, Ohio. It is part of the WPP Group.

P2PI MEMBERS

JOIN P2PI

You May Also Like

Cracking the Code on Incrementality in Brick-and-Mortar

Ghirardelli Enhances Digital Product Imagery with AI