Kellogg Magazine  |  Spring/Summer 2018



A Better Way to Understand How Customers Feel
About Your Brand

A new algorithm tracks people’s perceptions in real time via Twitter.

For decades, marketers have relied on surveys to gauge how cus-tomers perceive their brands. While this tried-and-true method does a good job of revealing how brands stack up against the competition on everything from health to luxuriousness, it is also time-consuming and labor-intensive. By the time you have survey results in your hand, they may already be out of date.
Illustration by Michael Morgenstern
Jennifer Cutler, an assistant professor of marketing at the Kel-logg School, thinks it may be time to send many surveys into a well-deserved retirement. Instead, she and a coauthor have de-veloped a real-time tool based on Twitter activity.

Cutler and Aron Culotta, of the Illinois Institute of Technology, created an approach that allows marketers to track in real time how their company compares to others for any attribute that interests them: in minutes, a marketer can know whether cus-tomers see Subaru as more or less eco-friendly than Toyota, a task that might previously have taken weeks or even months to complete. This is accomplished not by tracking what users are posting to Twitter, but rather whom they follow – an approach Cutler believes offers deeper and more nuanced insights into how companies are viewed.

“There’s a lot of excitement in the field of marketing about the potential to extract insights about consumers from these data, but there’s definitely been a struggle to figure out how to do that,” Cutler explains. Thus, a lot of the data remain untapped by mar-keters. Thanks to research like hers, however, “a lot of the barri-ers to entry and a lot of the obstacles to applying large-scale data mining for marketing insights are falling down.”

When marketers look to social media, they are often focused on what consumers are saying about their brands. Though Cutler believes text analysis has its place, there are serious limitations to relying on text alone. For example, although 20 percent of U.S. adults have Twitter accounts, fewer than half post actively.

“Although we talk about brand perceptions
specifically in this paper, the general idea of
looking to your users’ network connections
can be applied a lot of different ways.”
“Among those that write, very few are going to explicitly write about your brand, and even fewer still are going to write about the specific attributes of your brand that you are interested in measuring,” Cutler explains.

But consumers reveal a lot about themselves online, even when they say nothing at all.

These Twitter lurkers are following other users – companies, politicians, celebrities, friends – and making lists of accounts, organized by topic. Through lists, users can create their own cu-rated newsfeeds around topics of interest (“sports,” “science” or “politics”). And unless they have made their Twitter account pri-vate, all of this information is publicly available.

Across these many millions of user-curated lists, certain com-monalities begin to emerge. @ESPN, for instance, might appear on many user lists labeled “sports,” because users strongly asso-ciate it with that topic. Ditto @nytimes and “news.”

This is the basis of Cutler’s algorithm. The tool searches for ac-counts that appear on many lists labeled, for instance, “envi-ronment,” and narrows those accounts down to the strongest exemplars. For example, @SierraClub or @Greenpeace might be exemplar accounts for “environment.”

The algorithm then looks for overlap be-tween the followers of the exemplar ac-counts and the followers of a particular brand (say, Toyota). This information is used to compute a score that shows how the brand is associated with the attribute. Lower scores mean most customers do not associate the brand strongly with the attribute (say, Walmart and luxury); high-er scores indicate a stronger association.

To test the reliability of the method, the researchers compared their comput-er-generated results with traditional sur-vey results for 239 brands. In most cases, the survey results closely matched the results produced by the algorithm.

Overall, Cutler and Culotta found their tool provided a highly reliable measure of brand perceptions. And in contrast to the sluggish process of administering sur-veys, the algorithm can respond instantly to shifts in public perception or changes in a particular area of interest. “Anytime we want to run this model, we can just query again, and if there are new players in the field – new, trendy sustain-ability exemplars – then we’ll catch them with the new query,” Cutler says.

She hopes marketers will realize that “it’s important to consider your followers’ so-cial relationships and social networks on social media, not just what they say. What we’re showing here is that networks can provide a lot of extra information that is often missing in text.” It is an insight Cutler believes can be ap-plied much more broadly. “Although we talk about brand percep-tions specifically in this paper, the gener-al idea of looking to your users’ network connections can be applied a lot of dif-ferent ways,” she says. For example, sheis currently at work on a project that uses similar data-mining techniques to help marketers develop customer personas.

And she hopes as social-media data min-ing becomes more accessible to market-ers, it will allow them to gain insights into deeper and more abstract qualities of brand image. “As we develop these new techniques, it can start to open the door to new types of questions that marketers can ask that they haven’t been able to ask before,” she says.