Verena Schoenmueller Can Guess Your Political Preferences from Your Favorite Products

THE ASSISTANT PROFESSOR AT BOCCONI'S DEPARTMENT OF MARKETING OBTAINS INSIGHTS ON CONSUMERS BY OBSERVING THEIR BEHAVIOR ON THE INTERNET, BE IT ONLINE REVIEWS, FAKE NEWS SHARING OR TWITTER FOLLOWS

The Italian performer Giorgio Gaber once ironically and provocatively sang that the bathtub is rightist and the shower is leftist. It turns out that there may be a grain of truth in it. Verena Schoenmueller, an Assistant Professor at Bocconi’s Department of Marketing, observed which brands are followed on Twitter by people that follow Donald Trump, the Republican Party, Hillary Clinton, and the Democratic Party, and was thus able to create a measure of brand preference partisanship. «The favorite brands of the two groups are in fact different», she says, «and more so after the election». The analysis highlighted differences in sector interests (Trump’s supporters follow many finance, alcohol, and sport brands, while Clinton’s followers have a preference for tech companies), as well as in brands of the same category (they visit different retail stores, for example) – over and beyond differences in gender and location.
 
«As expected, media outlets differ considerably», Professor Schoenmueller says, «but not always in an obvious way. The New York Times, so often despised by Trump, results leaning only slightly to the left, for example. The print media outlet with the highest overlap to Clinton is The Atlantic; the online news source with the highest overlap to Trump the Drudge Report». If polarization in brand preferences is on the rise after Trump’s election, it seems due to Democrats more than to Republicans, the study suggests. «We interpret this result as a form of overcompensation after a threat to identity, as Liberals increasingly start to follow Democratic brands», Prof. Schoenmueller explains. She tackled the topic while being a post-doc researcher at Columbia University, NY. «The original idea was to draw a map of consumer preferences based on their behavior on social media, and when the elections came, we applied the idea to political brands».
 
A German, Verena Schoenmueller obtained her PhD in Marketing at the University of Basel, in Switzerland, after having studied in Germany and France. «I’ve done internships in the marketing department of companies, but I realized it wasn’t what I wanted to do». After her PhD, she was a postdoctoral research fellow at Columbia University with Oder Netzer. «There, I built on the topic of my PhD dissertation: online reviews». In a paper under revision in a top-tier journal, Prof. Schoenmueller analyzed the online reviews published on 25 platforms and found that reviews are predominantly very good, sometimes very bad, and only a few of them lie in the middle. «The reason is self-selection», she says, «we tend to rate only good things and sometimes bad ones, but mediocre things seem non-motivating. On the other hand, it implies that the average rating becomes less meaningful if there is high self-selection in online reviews, and that the number of reviews is a better indicator of quality in these situations».
 
After the experience at Columbia, Verena Schoenmueller decided to come back to Europe. «I appreciate the Italian culture and mentality and of course Italian food, so that restaurants are among my favorite places in Milan. I’m learning Italian, but I’m only a starter».
 
Another research interest of Schoenmueller’s is fake news. In a working paper, she analyzed fake news and fact check sharers’ emotions and personalities based on the tweets of users and found that there is much overlapping between fake news sharers and fact-checkers.  «Both of them are more angry and anxious than the average individual», she summarizes, «although they are often viewed to be on an opposite side of a continuum of fake news. Differences are that fact-checkers are more liberal, while fake news sharers are more conservative, and male. Interestingly, the users’ texts, which are generally more accessible on online platforms compared to information on demographics or political affiliation, show a similar predictive ability and can help to flag these users on online social media platforms. Something that seems of particular interest given the broad public discourse on fake news».
 
 
Find out more


Verena Schoenmueller, Oder Netzer, Florian Stahl, The Extreme Distribution of Online Reviews: Prevalence, Drivers and Implications, working paper.
 
Verena Schoenmueller, Oder Netzer, Florian Stahl, Polarized America: Going Beyond Political Partisanship to Preference Partisanship,working paper. 
 
Verena Schoenmueller, Gita Johar, Simon Blanchard, Fake News Sharers and News Fact-Checkers: Profiling and Targeting Beyond Demographics and Ideology, working paper.

by Fabio Todesco