The Tyranny of Most-Popular Lists

Easy rankings help us navigate a busy world by limiting and ordering our options. But the convenience comes at a cost.

Screen Shot 2013-06-17 at 11.53.50 AM.png


Let's talk about your next click.

At The Atlantic, we're always interested in understanding as much as we can about how you read us. What sort of headlines do the best traffic? How can we encourage you to finish a story and promote it on social media? Particularly challenging is getting readers to click on more Atlantic articles.

So we've put the question to our analytics team: When somebody clicks on an Atlantic article, where does she go next?

It turns out that, despite our best efforts to tease headlines at the top of the page and all along the right rail, the most popular "second click" is all the way toward the bottom. It's the "Most Popular" box listing the most-read articles on the site. Even after we redesigned our article pages and moved the location of the box, our data still showed that more people were clicking one of the ten articles in the "Most Popular" list than any other module on the article page.

Screen Shot 2013-06-17 at 8.20.11 AM.png

This is a little weird. The Most Popular Box is practically buried on our article pages below the last word most people read. It often reflects old articles pinging around email long after the news is fresh. Don't you guys want our new stuff? Apparently not.

Exploring the psychology of the digital marketplace
Read more

And yet, I'm as guilty as anybody of navigating by lists. I gravitate toward the "Most Popular" boxes on the New York Times, Slate, the New Yorker. I go to sites like Reddit and Digg specifically to learn what other people are clicking. I use Twitter as my homepage of news precisely because I'm more interested in what people are actually reading than in what institutional homepages are interested in selling.

I read (and, I'm guessing, you read) "most popular" boxes not only because we're savvy, but also because we're lazy. Where's the best stuff? is the question that motivates my Internet snooping. But that's actually a hard if not impossible question to answer. An easier question to answer is: What's everybody else reading? Hence, the power of the "Most Popular" box. Why drive yourself around a website when you can let the Internet's eyeballs drive for you?


The Internet didn't invent top-ten lists. In the markets for books, music, and movies, consumers navigate a complicated world of abundance by seeking lists and reviews to limit and order our options. But what effect does leading-by-lists have on our opinions of those books, music, movies, and experiences?

Sociologists Matt Salganik and Duncan Watts ran an experiment with music. Subjects received 48 songs listed in order of "most downloaded." They listened to the songs and downloaded their favorites.

There was a catch, naturally. One group saw the true rankings of the most-downloaded songs. The other group saw the rankings reversed, with the most popular song at the bottom and the least popular song at the top. It turned out that subjects with the reversed list downloaded the least popular song the most. The mistaken belief that a song was popular -- even if the song was demonstrably unpopular! -- made it most likely to be downloaded.

Essentially,  top-ten lists have a kind of placebo effect, the researchers concluded,. They don't just tell us what to read or listen to. They tell us what to like. And they make us think we like something just because we saw it next to the number 1.


If two sociologists can persuade their subjects that an unpopular song is a great song by putting a "1" next to it, you can imagine similar duplicity in the real world having significantly lager consequences. Yelp famously responded to a flood of fake reviews that were encouraging people to go to bad restaurants with false positive ratings. For the millions of Americans who orient their reading habits by top-ten lists, it means that buying books in bulk to force them onto a NYT bestseller list is the most effective form of advertising. Mobile apps have an incentive to cheat their way onto the front page of the App Store, since many app shoppers hardly make it to page two.

It's well understood that lists and rankings can be fixed. But Salganik and Watts' research makes a bigger claim: That fixed rankings can dupe us into liking things that we wouldn't have liked if they hadn't been ranked more highly. The placebo effect of most-popular lists suggests that better-reviewed meals might actually taste better; more-downloaded songs might actually sound better; articles with more Facebook likes might actually feel more delightful to read. When we outsource our navigation of the world to other peoples' opinions, we lose, in a small way, our ability to individually evaluate the quality of our experience.

Maybe that's a perfectly acceptable risk in exchange for the convenience of using simple lists to orient ourselves in a universe of stuff. Still it's sobering to think how much we're leaving to the "wisdom" of the crowd.