The “For you” feed is now ubiquitous across digital media platforms, including TikTok, X, Weibo, and Google News. Unlike the “Following” feed that you can meticulously curate by choosing just which accounts and outlets to follow, the “For you” feed seamlessly and mysteriously selects accounts, content, and ads for you using AI-informed algorithms.
Serving you up an endless stream of content based on your past behaviors—such as clicks, likes, views, scrolls, or even just lingering for a few seconds—within a frighteningly short amount of time, the “For you” feed just seems to get you. It’s akin to being served a continuous diet of only the most familiar and comforting foods.
But researchers are warning users against overindulging on the “comfort food” that is your “For you” feed.
Based on a massive field study that analyzed the behavior of millions of users of Weibo, they not only uncovered huge differences in user behavior between the “For you” and “Following” feeds, but also raised red flags about long-term effects on the psychology of users themselves.
“The focus of these platforms is on increasing users’ stickiness,” said Dr. Beibei Dong, an associate professor of marketing in the Lehigh University College of Business and one of the authors of the study. “Or in other words, to hook you.”
The prowess of algorithms on sites such as
YouTube,
TikTok, and
X in hooking and keeping users online is well-documented. However, Dong said, little academic research had previously been conducted to determine how different types of ads performed in the “For you” and “Following” feeds.
That’s both because big tech companies are loath to share proprietary information about their algorithms and because most universities lack the infrastructure required to conduct broad-scale studies, Dong said.
The researchers found their way in by partnering with a Chinese software company that was planning a massive ad campaign on Weibo. The partnership enabled the researchers to analyze data from 297 different ad versions that were served more than 300 million times to Weibo users—a scale unprecedented in U.S. academic research.
After the massive field study, researchers validated their results using laboratory techniques including eye-tracking on smaller groups of participants.
Their findings reiterated that AI can do a great job in getting to “know” users and serving them appealing content and ads. But there are differences in what is best for users, advertisers, and the platforms themselves.
The study, published in the
Journal of Marketing, found that native ads served in the “For you” feed had a 20.08% higher click-thru rate than ads served in the “Following” channel.
That’s a big difference. It’s easy to see why platforms are rapidly adopting AI-recommendation channels. After all, they get paid each time a user clicks an ad.
But advertisers may not necessarily be so thrilled with this result. If their aim is simply to increase brand recognition or visibility, then more clicks on their ads is a good thing. But if they’re looking to drive sales, then more clicks isn’t always better.
In fact, the researchers found that ads in the “For you” feed had a 15.6% lower conversion rate—that is, the percentage of clicks that lead to a purchase, or “conversion.”
So why are you so much more likely to click an ad in your “For” you feed but so much less likely to complete a purchase through it?
The answer, Dong says, relates to your level of “cognitive engagement.”
In the “Following” feed, content is organically served from accounts that you have chosen to follow and you’re more likely to find these accounts credible. As Dong said, “perceived control is higher in the subscription channel, and because of this, the level of engagement is higher.”
When you’re investing more cognitive effort in consuming the content, you’re less likely to distractedly click on an intrusive ad. With this higher level of cognitive investment, you’re more likely to click an ad because you’re actually interested in purchasing it.
Advertisers can capitalize on this phenomenon, Dong said, by crafting ads that optimally use informational or emotional appeals and direct or indirect links.
Informational appeals focus on presenting factual information about products, while emotional appeals emphasize the feelings associated with the product experience.
Direct links take users straight to a purchase screen (e.g. “Buy now”), while indirect links lead to other information (e.g. “Click to learn more”).
Within the “Following” feed, to drive more clicks, advertisers should use emotional ads with indirect links, researchers found. To drive conversions, they should use informational ads with indirect links.
Conversely, in the “For you” feed, users tend to slip into a state of lower cognitive engagement. With the lower investment of cognitive effort required, ads begin to feel less intrusive and annoying, and users are more likely to click on them. However, they’re less likely to exert the cognitive effort required to go through with a purchase based on that click.
To optimize their ad spend in the “For you” feed, marketers should use emotional ads with indirect links, which decreased click-thru rates by more than 20% and increased conversions by almost 75%.
To drive more clicks in the “For you” feed, marketers should use informational ads with direct links, which increased click-thru rates by more than 31% but decreased conversion by more than 25%.
Of course, it’s possible that what’s good for platforms and what’s good for advertisers is not what’s good for users. That can be the case when consuming AI-recommended content, researchers warned.
Over time, relying too much on the algorithm to curate your content diet can have long-term cognitive effects. In addition to the temporary “low engagement mode” that leads you to click on more ads, prolonged existence can create a “filter bubble” in which users receive only content that perfectly aligns with what they have already consumed.
As other studies have shown, this can lead users down “
rabbit holes” of extreme content, conspiracy theories, and deepened political and cultural divides. Dong cautioned against the effects of the “filter bubble” on users’ critical thinking skills, ability to process new information, and creativity and learning ability.
So while it’s ok to indulge in your favorite digital comfort foods, they should not compromise your entire media diet.
“You can exert a two-way influence on what AI learns about you,” Dong said. “Be more selective in interacting with subscription channels. Just like eating a well-rounded diet is good for you, intentionally seeking out different ideas and perspectives in the media can help fight the negative effects of the filter bubble.”
Photo: DIW-AIgen
Read next:
Here’s What Happens on the Internet Every Minute
by Irfan Ahmad via Digital Information World
