How Television Advertising Quality Affected McDonald’s Sales Growth Over Six Years
The current research used McDonald’s data to explore the relationship between advertising quality and sales growth. Based on a 6.5-year dataset involving more than 180,000 consumer interviews, the researchers found that nearly half of McDonald’s sales growth could be explained by variables related to advertising quality. Specifically, the study found that factors such as sales momentum, the introduction of calorie content on to quick-service restaurant (QSR, i.e., fast-food) menus, and variables related to key research metrics—and, importantly, the right messaging strategy—can be effective in linking advertising to variation in sales.
An ongoing challenge to advertising researchers is how to validate predictions about real-world sales performance based on pre-testing metrics. Such validations are important for three reasons:
- quantifying the contribution advertising can make to return on investment (ROI),
- confirming the general construct of how advertising “quality” works can be implemented as a quality-control step in the creative process— something once lamented as a “dream that will never come to be” (Metzger, 2013), and
- providing a baseline for distinguishing between what is known and not known about how advertising works is a critical element in evaluating improvements in the research process.
Much previous research has found that the quality of advertising matters in terms of perceived brand perceptions and marketing efforts (Dahlén, Rosengren, and Torn, 2008) and that it drives mind-set metrics like cognition, affect, and experience that can be linked to sales performance (Bruce, Peters, and Naik, 2012). To the current authors’ knowledge, however, a large dataset analysis never has been performed to link advertising quality and sales.
The current study explored the relationship between the quality of McDonald’s U.S. television advertising creative product and publicly reported McDonald’s sales figures. Advertising quality was measured by a testing system created by Ameritest, the international marketing-research firm based in Albuquerque, NM.
Given the size and extended period of the dataset, the authors believe that the current analysis offers a “big-data” approach to validation. They also believe that they can explain nearly half of the sales growth reported by McDonald’s during this 6.5-year period by examining four variables:
- sales momentum;
- advertising-quality metrics;
- calorie variables (beginning when McDonald’s began offering calorie counts on its menus); and
- negative advertising quality (advertising that failed to communicate its intended message).
Importantly, each variable relates directly to the quality of the advertising creative work. Moreover, the size of the advertising investment was not a consideration in the current study; McDonald’s spending data is proprietary information, and the research team did not have access to it. One benefit of this necessary omission is that it sets this study apart from other marketing-mix models that often tend to be driven heavily by media spend. Another benefit was that the researchers could establish the importance of creative quality more clearly without regard to advertising expenditure.
MCDONALD’S METRICS AND DATA
For a variety of reasons, the research team found McDonald’s to be a particularly apt subject for a research-validation exercise.
Given the size and extended period of the dataset, the authors believe that the current analysis offers a “big-data” approach to validation.
For one thing, as quick-service restaurants (QSR) tend to produce much advertising in short periods of time, a television advertising investment likely would have a great influence on sales.
In a typical year, the top 18 QSR brands produce and place more than 300 thirty-second advertisements on national television. As the category leader, McDonald’s airs approximately 20 percent of the category’s distinct television spots each year. With this kind of investment, McDonald’s management certainly would expect to see a strong relationship between advertising and sales.
Moreover, because of its fast-tempo advertising strategy, McDonald’s changes out its creative product monthly, with four to five new commercials airing each month. As a result, the time frame for seeing a direct advertising-to-sales effect should offer a fairly narrow window for analysis. This pace allowed the authors of this study to have more confidence in the variables they chose to use. In fact, due to the rate at which advertising changes in the category, sales theoretically should change quickly in response to specific advertisements.
Regular validation exercises are standard for most major pre-testing clients. The results of such work, however, often are proprietary. For the current study, the authors used either publicly available information or an Ameritest database of syndicated creative scores.
The sales data used in the current analysis are the monthly change in same-store sales compared to the prior year for McDonald’s U.S. operations—information reported to Wall Street and obtained from Morgan Stanley by the research team.
The data covers 6.5 years, from January 2007 through May 2013. To test the stability of the model, the authors first studied the 5-year period that ended in 2012. They then updated that analysis with data from the sixth year. Essentially, the findings were similar for both time periods.
The sample of advertisements used in the study represented a virtual “census” of all 30-second QSR commercials aired in the United States in the past 6.5 years—or almost 2,000 commercials—with 180,000 consumer interviews. Of these commercials, 441 were for McDonald’s, and 1,533 were for McDonald’s competitors.
The creative metrics used were collected continuously using a standard methodology developed by Ameritest over a significant period of time and among a controlled sample of consumers for a large number of advertisements.
A model predicting McDonald’s sales was built using multiple regressions. To keep the model parsimonious, only four variables were used. The model related sales for any given month to the performance of commercials airing that month. This four-variable model explains almost half of the variation in sales growth, with an R-squared of 48 percent.
- The variables are defined as follows (Y— Monthly change in same-store sales):
- X1: A “momentum” variable—the average of the prior three months’ sales growth—that accounted for longer-term effects of advertising
- X2: An advertising-quality metric comprised of executional performance metrics and communication metrics
- X3: A “calorie” variable—in this case, a dummy variable set to 0 and to 1—that allowed the research team to account for the period beginning in late 2012, when McDonald’s began offering calorie counts for its products on its menus
- X4: A negative-advertising quality variable that identified any advertisements that had failed to communicate a relevant strategic message to consumers other than the default consumer inference of “convenience.”
Y = 0.50X1 + 0.39X2 + 0.20X3 – 0.39X4 – 0.826
Further Explanation of the variables
The momentum variable (X1) was designed to reflect the idea that month-to-month changes in sales were not independent of one another but rather built on one another over a period of time.
The advertising quality variable (X2) was a composite index:
Ad Quality = Execution × Strategy = Creative Idea × Relevant Message = (Ameritest Performance Index)× Messaging
The Ameritest Performance Index (API) consisted of “Attention,” “Branding,” and “Motivation” (intent to visit restaurant), which are standard pretest “report-card” measures.
The messaging metric (X3) is the top-rated message from the group of ten fast-food messages collected in the brand ratings. To calculate the composite creative index score for each month, the researchers examined the messaging of each advertisement and included only “effective” advertisements that scored at (or above) average (index greater than 100) on at least one strategic message.
The API scores for these qualifying advertisements were multiplied by the top message rating scored by that advertisement. From those data, the authors were able to calculate an average creative index score for the number of effective advertisements running that month.
The negative-advertising quality variable (X4) was calculated in a way that was analogous to X2, except that only advertisements that communicated “convenience” as the top-rated message were used in the average for each month; advertisements with low and high API scores were included.
The Research Model
The momentum variable (X1) is the most important of the four variables examined in the study and accounts for about a third of sales growth, with an R-squared of 34 percent.
The advertising quality variable (X2) is the next most important, with an R-squared of 9 percent. Of course, this part of the analysis accounts only for advertising’s contribution to short-term sales and is not tied to long-term brand-building effects that may result from advertising used to establish a brand positioning and image. These types of effects would be contained in the momentum variable.
The advertising quality variable (X4) that accounts for ineffective advertisements that fail to contribute a strategic message also enters the model in a negative way—with a negative beta coefficient—to increase the predictive power of the model by another 2 percent.
Finally, the dummy variable marking McDonald’s decision to communicate calorie content (X3) adds 3 percent to the model.
The Importance of Strategic Communication
In one seminal analysis, communication was found to occur only when a particular message is selected from a set of possible messages that must be known to both sender and receiver (Shannon and Weaver, 1949). For the QSR industry, these messages include such ideas as taste, variety, health, and a family-friendly atmosphere, and each message can vary in its creative expression and degree, (e.g., Dahlén et al., 2008).
By contrast, Subway’s strategy is quite different: Its messaging has focused entirely on health claims, as the QSR has attempted to position its brand against the entire fast-food category, which often is perceived as unhealthy.
The presence of a strategic message by itself is not enough to explain advertising’s impact on McDonald’s sales growth. In combination, however, with industry measures of executional strength—“attention,” “branding,” and “motivation”—it is possible to assess how the quality of the creative product in advertising may drive sales.
The Impact of a Failed-Message Strategy
One way that advertising can fail to be effective is by not communicating a message that supports the strategic positioning of the brand. As the authors of this study observed in certain McDonald’s messaging strategies, this kind of failure may do damage that extends beyond the lost investment of misplaced media dollars. Creative work that strays too far from what the brand stands for in the mind of the consumer can blur the brand positioning and damage the brand image.
An example: A McDonald’s advertisement that reminds consumers of only the convenience of eating at McDonald’s. In practice, the QSR never would run advertising with the primary messaging intent of the categorical attribute— “convenience”—which so often has negative connotations of poor quality and poor nutrition throughout the competitive set.
When looking at the distribution of McDonald’s advertising by messaging, the authors of the current study found that 60 percent of commercials conveyed a primary message of one or more strategic benefits (i.e., taste or variety). However, for 30 percent of McDonald’s advertisements, the primary or only message communicated by the advertisement was “convenient.” In the absence of any stronger conveyed benefit, “convenience” becomes the default message taken away by the advertising. The remaining 10 percent conveyed none of these strategic QSR messages.
As the meaning of a McDonald’s advertisement defaulted to convenience—or to nothing at all—the quality rating declined accordingly. More specifically, in looking at the correlation with “motivation” scores, when convenience was the only message of an advertisement, the correlation was negative (–21 percent) but could be a positive (+26 percent) when convenience was couched with other relevant messages.
Thus, in the current model, the authors found evidence not only that the quality of the advertising creative product was an important variable in driving sales performance but that unfocused advertising actually could hurt sales.
Additionally, the authors discovered that finding a strong strategic message was important to ward off competitive attacks. In months when McDonald’s communicated a stronger strategic message than its competition, sales tended to increase. During months when competitive spots more effectively communicated a strategic brand message, McDonald’s sales growth stagnated or even declined.
This finding highlighted the need to be aware of the competitive set in the marketplace. In fact, when McDonald’s had an execution in the top-three performers among QSR advertisements for any given month, average sales were half a percent higher than when McDonald’s did not have a top-three advertisement. And that half-percent may amount to hundreds of millions of dollars, underscoring how important it is for McDonald’s to know where it stands versus the kind of advertising its competition is running.
Long-Term Advertising Effects
Judging from the betas in the model, the authors of the current study found that the momentum variable (X1), which described effects that persisted longer than a single month, was most important. This variable summarized a variety of marketing inputs, including changes in service operations, changes in product offerings and, of course, the longer-term contributions of advertising to the McDonald’s brand.
The link between memory and experience is clear but multi-faceted (Kahneman, 2011). A consumer’s “brand memory” of McDonald’s, therefore, is complex, composed of all past experiences (real or imagined) that the consumer has had with McDonald’s, including engagement with advertising.
To fully understand how advertising drives sales, it is essential to understand that the “brand” is nothing more than a memory. An effective McDonald’s advertisement transcends the simple stimulus response of sales promotions and thus can be expected to have an effect on sales for some time after it has aired.
To demonstrate how advertising persists in memory, the authors of the current study performed a simple experiment with five highly effective, attention-getting fast-food commercials. Tested among three different consumer samples per advertisement, the authors measured memory—frame by frame—20 minutes after commercial exposure, 24 hours after commercial exposure, and seven days after exposure.
Not surprisingly, the greater the amount of time consumers had to forget, the less they remembered. While an average of 77 percent of images from these commercials were remembered after 20 minutes, after a full day, only 62 percent were remembered. After a week, consumers remembered 52 percent of the images from the commercial.
Importantly, when the researchers looked across the five commercials, they saw similar patterns with the rhythmic peaks and valleys of the memory map. The images that were peaks after 20 minutes remained peaks after longer time periods. In other words, the images that were key to predicting breakthrough scores or commercial engagement were the same images that lodged most strongly in long-term memory to build a brand’s image.
As these images persist over time in consumer memory, they likely form the emotional basis of consumer loyalty to complement a more rational basis for loyalty (Reynolds and Phillips, 2005). And that emotional bond suggests the mechanism for advertising’s long-term contribution to the momentum variable in the current model.
IMPLICATIONS AND CONCLUSION
Marketing managers seek to understand the relationship between advertising quality and sales growth. This McDonald’s analysis—with a 48 percent R2 on sales growth by using advertising quality metrics—demonstrates that this link is not only possible but highly important to a full marketing-mix model.
The authors of the current study believe the link between the quality of creative work in advertising and short-term sales growth is clear. This highlights the ever-evolving need to generate strong insights about creative quality, so a marketer can both understand the “whys” of his or her advertising and take this into account when making a sales forecast. Linking strong creative quality with established marketing-mix model principles, such as sales momentum and new category information (e.g., as in the case of McDonald’s calorie content communication), should give the marketing manager a clearer picture of the marketplace and expected results.
To improve understanding of the link between advertising and sales growth, marketing managers should take note of the following conclusions the authors have drawn from this study:
• Independent of media-spending levels, the quality of the advertising creative product is a major factor driving sales response. Marketing-mix models attempting to quantify ROI are incomplete if they do not include a quality variable for the creative work.
• Relative performance of advertising versus the competitive set is important—one of the “fundamental questions facing brand marketers” (Reynolds and Phillips, 2005). The authors believe, therefore, that competitive testing and awareness is a necessary component in the analysis of fluctuations in sales figures. Such analysis can lead to the identification of areas where the brand needs to position to fend off competitive attacks.
• Creative “quality” is a composite of message communication and executional variables. This distinction parallels two common ways of thinking about a brand: “positioning” versus “brand image.” A brand manager’s structural composition of an advertisement and messaging strategy, therefore, needs these two elements to be linked at optimum levels to ensure that both positioning and brand image can be understood completely by the consumer. Also, one cannot be traded off for the other; communicating a relevant message is as essential to the advertising story as the executional pieces needed to hold that message together.
• Longer-term effects of advertising on sales need to be better understood and quantified. Much previous research has focused on “emotional engagement.” Research suppliers consistently have introduced new products and services—among them biometrics and facial response techniques—that measure the emotional response of the consumer.
Even in the face of such metrics, the authors of the current study believe that the importance of “brand memory”—as it relates to loyalty and consumer trends—is a missing part of the emotional equation. Future research should shift attention to the study of how these emotional engagement strategies connect with advertisements to create long-term brand memories.
Charles Young is founder and ceo of Ameritest, a global research firm, based in Albuquerque, NM. Young invented the firm’s Ameritest’s Picture Sorts technique that analyzes visual components of advertising. Previously, he was research partner for Euro/Tatham and a new-product consultant for Leo Burnett. In 2004, Young won the Advertising Research Foundation’s Grand Ogilvy Award for a case study conducted with IBM. He is the author of The Advertising Research Handbook (Ideas in Flight Publishing, Seattle, WA, 2008).
Adam Page is an associate research and analytics director at Ameritest, in which capacity he has developed statistical models for a number of Ameritest clients.
BRUCE, N. I., K. PETERS, and P. A. NAIK. “How Advertising Grows Sales.” Journal of Marketing Research 49 (2012): 793–806.
DAHLEN, M., S. ROSENGREN, and F. TORN. “Advertising Creativity Matters.” Journal of Advertising Research 48, 3 (2008): 392–403.
KAHNEMAN, d. Thinking, Fast and Slow. New York: Farrar, Straus and Giroux, 2011.
METZGER, G. “In Search of Advertising ROI: The Impossible Dream versus ‘Bounded Rationality.’” Journal of Advertising Research 53, 3 (2013): 251–253.
REYNOLDS, T. J., and C. B. PHILLIPS. “In Search of True Brand Equity Metrics: All Market Share Ain’t Created Equal.” Journal of Advertising Research 45 (2005): 171–186.
SHANNON, C. E., and W. WEAVER. The Mathematical Theory of Communication. Urbana, IL: University of Illinois Press, 1949.