The murder of George Floyd and the wave of protests that followed it led many companies to feel they should issue statements of support for racial justice. But not all companies responded the same way.
Chicago Booth's Pradeep K. Chintagunta and his coauthors analyzed the emotionality of tweets about five different companies to see how their responses to the Black Lives Matter movement affected public sentiment. They find that the companies in their analysis ultimately wound up in a better position, in terms of the emotional content of tweets about them, but few of them succeeded on their first attempt at a response. To be positively received, the analysis suggests, corporate responses to social issues need to include a significant commitment to the cause.
Narrator: Thus far, 2020 has been a challenging year, to say the least. The COVID-19 pandemic, economic collapse across countries,autonomy protests in Hong Kong, protests of police brutality and racism in the United States, and the massive chemical explosion in Beirut are a few of the major events and conflicts that have dominated public consciousness of late. So it’s no surprise that in addition to being turbulent times, these are emotional times, and the way we talk on social media has shown to reflect that.
Chicago Booth’s Pradeep Chintagunta and his coauthors have been testing the positive emotionality of tweets and how they have changed from January to May 2020. And it may not surprise most of us, but it has declined.
Pradeep K. Chintagunta: There are many ways of thinking about positive emotionality. If you go back prior to when we could do computerized text analysis, they would look at the text of a document and then try to count up the number of words which were positively emotioned and the number of words which were negatively emotioned. Of course, that’s very hard to do when you have lots of texts. And so these days, with the presence of computerized text analysis, what you can do is to use software that’ll help you do this.
So in our case, we used something called LIWC, which actually stands for Linguistic Inquiry and Word Count. And what it does is it tells you the proportion of words in a document which have a positive emotion—things like “happy,” etc.— versus the words that are negatively emotioned— things like “sad,” “unhappy,” and the like. Using this software, we are able to compute the percentage of wordsin a given document or tweet which essentially are either negatively emotioned or positively emotioned.
Narrator: On May 25th of this year, George Floyd was killed by Minneapolis police officer Derek Chauvin. The event sparked a highly charged and increasingly polarized racial-equality movement in the US.
Many activists have called out firms on Twitter, pressuring them to take a stand against racial inequality, putting these companies in unfamiliar territory.
Pradeep K. Chintagunta: Companies try to shy away from things which are very controversial, largely because there’s a fairly large degree of heterogeneity in their customers, in their employees, and other stakeholders. So taking positions which tend to be charged either politically or socially was always somewhat tricky.
Some companies clearly have responded. They have obviously responded in different ways. I think a lot of companies that have responded have tended to be much more consumer facing, but there are many other companies which have chosen not to do a whole lot.
Narrator: The researchers selected five customer-facing companies: Airbnb, Apple, Amazon, Netflix, and Pepsi. Using a control, they compared the emotionality of tweets about them a week before and a week after they released statements supporting racial equality.The data showed some interesting results.
Pradeep K. Chintagunta: Most of these companies came out on the positive dimension. And what I mean by that is that the emotionality of the tweets tended to be more positive than they were before the announcement. But it didn’t happen instantaneously. So it depended a lot on the nature of the announcement that was made.
Oftentimes, when companies simply made an announcement without really signaling any commitment behind the announcement—and by commitment, I mean basically financial commitment—it turned out that the emotionality of the tweets actually went more negative than it went positive. And it’s only when these companies decided to commit real resources to these causes that you find that there’s a turn in the emotionality, and the positive emotionality of the tweets actually goes up.
Narrator: Netflix is a good example of a company that faced negative emotionality after their first statement because they failed to make any kind of financial commitment for racial justice. Their second announcement, which had a financial commitment, was also received poorly. It seemed the public did not view that commitment as being large enough to be convincing. So they released a third announcement, where they then committed as much as $100 million to support black communities in the US. It was only at that point that the emotionality of tweets about Netflix became more positive.
After the Netflix episode, many companies internally debated how they could respond, or even if they should.
Pradeep K. Chintagunta: A lot depends upon the company itself: what its position is, what its values are. And if these are issues that resonate very closely with the values of the company, then it makes sense for them to voice their opinions.
The other thing to keep in mind is that the customers of a lot of these companies now tend to demand this of the companies that they tend to transact with because they would like to associate themselves with products or product images which are more consistent with the way they feel. And especially with millennial customers, you see that these customers really like to understand the story behind the brand. What is it that the brand actually stands for? What is the positioning behind the brand?
So companies respond very differently. For example, Ben & Jerry’s is a good example of a company which is very closely associated with social causes. And so they have played a very active role in a number of causes like LGBTQ issues. Those kinds of issues are very much in their DNA. And I think part of that is the expectation from their own customers and other stakeholders for them to actually say something when an issue like this crops up.
An expert panel discusses what increasingly large amounts of quantitative data mean for companies, policymakers, and organizations.
What Are the Limits of Big Data?
Testing assumptions in the field can avoid costly errors in decision-making.
You Should Run More Experiments
Evaluating the performance of machine-learning tools isn’t always easily done.
How (In)accurate Is Machine Learning?Your Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.