Wondering How Customers Feel about Your Brand?
Skip to content
Marketing May 2, 2018

Won­der­ing How Cus­tomers Feel about Your Brand?

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

Consumer brand perception on social media

Michael Morgenstern

Based on the research of

Aron Culotta

Jennifer Cutler

For decades, mar­keters have relied on sur­veys to gauge how cus­tomers 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 lux­u­ry, 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 By the time you have sur­vey results in your hand, they may already be out of date.

Add Insight
to your inbox.

We’ll send you one email a week with content you actually want to read, curated by the Insight team.

Jen­nifer Cut­ler, an assis­tant pro­fes­sor of mar­ket­ing at the Kel­logg School, thinks it may be time to send many sur­veys into a well-deserved retire­ment. Instead, she and a coau­thor have devel­oped a real-time tool based on Twit­ter activity.

Cut­ler and coau­thor Aron Culot­ta of the Illi­nois Insti­tute of Tech­nol­o­gy have 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 cus­tomers see Tes­la as more or less lux­u­ri­ous than Porsche, a task that pre­vi­ous­ly might 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. 

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

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, much of that data remains untapped by mar­keters. Thanks to research like hers, how­ev­er, a lot of the bar­ri­ers to entry and a lot of the obsta­cles to apply­ing large-scale data min­ing for mar­ket­ing insights are falling down.” 

The Pow­er of Social Media Data Min­ing

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 draw­backs to rely­ing on text alone. For exam­ple, although 20 per­cent of US adults have Twit­ter accounts, few­er than half post actively. 

Among those that write, very few are going to write about a brand, and even few­er still are going to write about your brand,” 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 curat­ed news­feeds around top­ics of inter­est (“sports,” sci­ence,” or pol­i­tics”). And unless they have made their Twit­ter account pri­vate, all of this infor­ma­tion is pub­licly available. 

Across these many mil­lions of user-curat­ed lists, cer­tain com­mon­al­i­ties begin to emerge. @ESPN, for instance, might appear on many user lists labeled sports” because users strong­ly asso­ciate it with that top­ic. Dit­to @DogRates and cute,” or @nytimes and news.”

This is the basis of Cutler’s algo­rithm, which iden­ti­fies exem­plary accounts for par­tic­u­lar top­ics. The tool search­es for accounts that appear on many lists labeled, for instance, envi­ron­ment,” and nar­rows those accounts down to the strongest exem­plars. In the envi­ron­ment” exam­ple, @SierraClub or @Greenpeace might be exem­plary accounts. 

The algo­rithm then looks for over­lap between the fol­low­ers of the exem­plary accounts (@Greenpeace) and the fol­low­ers of a par­tic­u­lar brand (say, Toy­ota Prius). This infor­ma­tion is used to com­pute a score between zero and one 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); high­er scores indi­cate a stronger asso­ci­a­tion (Toy­ota Prius and the environment). 

To test the reli­a­bil­i­ty of the method, the researchers com­pared their com­put­er-gen­er­at­ed results with tra­di­tion­al sur­vey results for 239 brands. The researchers recruit­ed sur­vey par­tic­i­pants online and asked them to rank each brand from one to five accord­ing to how strong­ly they asso­ci­at­ed it with one of three attrib­ut­es: eco-friend­li­ness, lux­u­ry, and nutri­tion. They found that in most cas­es, the sur­vey results close­ly matched the results pro­duced by the algorithm. 

One inter­est­ing excep­tion: sur­vey par­tic­i­pants rat­ed Lam­borgh­i­ni high­er on lux­u­ry than the algo­rithm did. It was an intrigu­ing anom­aly, con­sid­er­ing the brand’s rep­u­ta­tion and the eye-pop­ping cost of its cars. So they looked more close­ly at Lamborghini’s Twit­ter account to fig­ure out why the algo­rithm might have faltered. 

In con­trast to oth­er com­pa­nies, which use Twit­ter to inter­act with cus­tomers or share infor­ma­tion about their prod­uct, Lamborghini’s account was a more gen­er­al news feed on car technology. 

We think that it’s quite pos­si­ble that peo­ple were fol­low­ing Lam­borgh­i­ni and news feed’-type brands for sys­tem­at­i­cal­ly dif­fer­ent rea­sons than why they might fol­low oth­er brands,” Cut­ler explains. 

Why con­sumers fol­low par­tic­u­lar brands, and how brands use social media to achieve a vari­ety of strate­gic aims, is some­thing Cut­ler hopes to inves­ti­gate more deeply in future research. 

Future Appli­ca­tions

Over­all, how­ev­er, Cut­ler and Culot­ta found their tool pro­vid­ed a high­ly reli­able mea­sure of brand per­cep­tion. And in con­trast to the slug­gish process of admin­is­ter­ing sur­veys, the algo­rithm can respond quick­ly 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 sus­tain­abil­i­ty 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’ social 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 applied much more broadly. 

Although we talk about brand per­cep­tion specif­i­cal­ly in this paper, the gen­er­al idea of look­ing to your users’ net­work con­nec­tions can be applied a lot of dif­fer­ent ways,” she says. For exam­ple, she is 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 min­ing becomes more acces­si­ble to mar­keters, 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.

About the Writer

Susie Allen is a freelance writer in Chicago.

About the Research

Culotta, Aron and Jennifer Cutler. 2016. “Mining Brand Perceptions from Twitter Social Networks.” Marketing Science 35(3):343–362.

Read the original

Suggested For You

Most Popular


How Are Black – White Bira­cial Peo­ple Per­ceived in Terms of Race?

Under­stand­ing the answer — and why black and white Amer­i­cans’ respons­es may dif­fer — is increas­ing­ly impor­tant in a mul­tira­cial society.


Pod­cast: Our Most Pop­u­lar Advice on Advanc­ing Your Career

Here’s how to con­nect with head­hunters, deliv­er with data, and ensure you don’t plateau professionally.

Most Popular Podcasts


Pod­cast: Our Most Pop­u­lar Advice on Improv­ing Rela­tion­ships with Colleagues

Cowork­ers can make us crazy. Here’s how to han­dle tough situations.

Social Impact

Pod­cast: How You and Your Com­pa­ny Can Lend Exper­tise to a Non­prof­it in Need

Plus: Four ques­tions to con­sid­er before becom­ing a social-impact entrepreneur.


Pod­cast: Attract Rock­star Employ­ees — or Devel­op Your Own

Find­ing and nur­tur­ing high per­form­ers isn’t easy, but it pays off.


Pod­cast: How Music Can Change Our Mood

A Broad­way song­writer and a mar­ket­ing pro­fes­sor dis­cuss the con­nec­tion between our favorite tunes and how they make us feel.