:format(webp)/cdn.vox-cdn.com/uploads/chorus_image/image/63873552/GettyImages_900224666.0.jpg)
In April, the New York Times announced that it had been using machine learning and surveys to predict the different emotions felt by readers of various stories.
The newspaper has been collecting this information for a year, and has created a list of 30 commonly experienced emotions — 18 of which are available for sale to advertisers, as reported this week by AdWeek. (Meaning an advertiser for, say, analog film could opt to have ads placed only on articles that induce nostalgia, while an advertiser for something I can’t even imagine could opt to have ads placed only on articles that induce fear.)
Those emotions, ranked in descending order by how much I would want to be targeted based on them, are as follows:
- Bored
- Amused
- Indulgent
- Optimistic
- Interested
- Adventurous
- Self-confident
- Inspired
- Informed
- Competitive
- Happy
- Nostalgic
- Fearful
- Hateful
- Hopeful
- Sad
- In the mood to spend
- Feeling love
This list was difficult to arrange. For example, it took three full minutes to decide that it is more painful to think about ads being targeted to me when I am feeling “love” than it is to imagine that “in the mood to spend” is a human emotion, never mind one that I am capable of feeling in a quantifiable way. According to AdWeek, “sadness” ads are popular with “socially responsible brands targeting women.” So, I look forward to that.
USA Today has had an emotion-based ad offering since 2017, and the Daily Beast is also testing something similar. Kelly Andresen, head of USA Today’s content studio, told AdWeek that the company had initially assumed that readers would only read “uplifting, positive news,” and was surprised to learn that people actually read everything. “This gives us an opportunity to target them when they are in the right contextual environment and the right state of mind,” Andresen explained.
The idea is that emotion is just another metric. Companies are increasingly skittish about buying ads paired with content that doesn’t match their brand “story” or pinpoint a highly specific target audience with just the right “values.” Having a list of emotions in front of them means they can — within the limits of publishers’ machine-learning capabilities — ensure that their brand is experienced in exactly the right context.
Something I think about a lot is a negative review of Inside Out, Pixar’s 2015 movie about anthropomorphized emotions, written by New Yorker film critic Richard Brody. He called the movie a “feature-length sales pitch — or, worse, an indoctrination — to mold kids into beings as artificial and uniform as those created by computer graphics in the movie.” While this argument is perhaps a little dramatic, Brody has a point. “In lieu of the mysteries and wonders of life, instead of big dreams and big fears,” he writes, “Inside Out offers problems to be solved, a narrow range of a narrow life of narrow prospects and narrow experiences, narrow fantasies and narrow desires[.]”
Emotion-based advertising can seem similarly simplistic, limiting, and downright weird.
“Brands don’t necessarily want 5,000 emotions to choose from,” Allison Murphy, the Times’ senior vice president of ad innovation told AdWeek, giving the example that ad-buyers do not care about the difference between happiness and ecstasy.
Again, it’s not clear which is more off-putting: having our emotions sold out to the benefit of brands, or being plopped into large buckets of nonspecific emotions that seem useless or inhumane. It would likely be better not to be manipulated; it would also likely be better not to see capitalism further encourage us to reduce other people to insufficient categorical tags. The question becomes whether emotion-based advertising should get better or just not exist.
Sign up for The Goods’ newsletter. Twice a week, we’ll send you the best Goods stories exploring what we buy, why we buy it, and why it matters.
