Couch potato: How Netflix’s algorithm affects what you watch — or doesn’t

A person wrapped in a red blanket sits in a dark room watching Netflix on their TV
Graphic by Annie Wu

Kids these days have it easy. I remember when the only non-word of mouth way to discover a new TV show that fit your interests was a website. 

You typed in whatever TV show you had just finished, like “Lost” or “Pretty Little Liars,” and then the algorithm would feed you a list of similar-ish shows. Watch “The Vampire Diaries! Or Fringe!” it would say. 

Now, streaming services like Netflix do that work, so you don’t have to leave your tab to get TV show recommendations customized to your unique watching history. As someone who really liked “Lost” in middle school, I would’ve been recommended “Black Mirror,” because data shows that people who like one of those shows tend to like the other. 

Netflix’s algorithm subliminally shapes how we watch TV shows through the content it suggests, from the length of the shows, to how award-winning they are, to the actors that are featured. Pundits enjoy criticizing people for living in bubbles, in which the content they consume and opinions they have are recycled in a narrow feedback loop. But nowhere is the banal presence of bubbles more apparent than in Netflix’s user interface. 

If shows are only recommended on the basis that they’re tangential to previously watched shows, opportunities to branch out and explore different genres are limited. Netflix’s algorithm essentially holds a monopoly on the main page. For example, those little headlines, like “trending now,” imply neutrality, but the TV shows that appear on “trending now” are completely different for each user — the recommendations are based on your watching history. 

All this is to say, I’m skeptical of Netflix. To suss out the extent of the algorithm’s influence, I did a deep dive on my own Netflix account compared to the account of my perfect opposite: a middle-aged man. (This other Netflix account may or may not be a family member’s, so it’s ethically sourced, don’t worry.)

Let’s start with my watching history. The comedies I’ve watched on Netflix include “Arrested Development,” “Parks and Recreation,” “New Girl,” “Crashing,” “Glee,” “Derry Girls” and “Easy.” For dramas, I have “Unbelievable,” “The Politician,” “You” and “Bodyguard.” 

My recommended categories are, in order: Award-Winning British TV Dramas, Emotional TV Dramas, Award-Winning Binge Worthy Crime Drama, Award Winning TV shows, three more categories with a synonym for Award-Winning, and Watch in One Weekend. 

The giant advertisement at the top of my main page is for “Schitt’s Creek,” a show I don’t love (it was a bit too cheesy for me) but many of my friends do. 

Conclusion: Netflix thinks I’m a cultured woman, one with no weekend plans. I don’t love the recommendations, but I won’t condemn Netflix for suggesting them.

Now for the middle-aged man. 

The first thing I notice about this account is that it’s darker. All of the TV show widgets (those slides that play a preview if you hover your mouse over them) are in black, perhaps with a splash of dark blue or red. The advertisement at the top is for “Peaky Blinders,” a show that — from what I can surmise from the pop-up — is about an angsty man living in 1920s industrial London who loves fedoras. 

The list of Netflix Originals recommended is also different than mine. While my top recommendations were for fun shows about murder such as “The Politician” and “You,” this Netflix account was recommended bleak shows about murder such as “Mindhunter,” “Ozark” and “The Spy.” 

At first, the different widget pictures seemed arbitrary, based on color more than anything else. For example, when I was recommended “Dexter,” it pictured Michael C. Hall, who plays the titular character, leaning on a couch. On this Netflix account, it’s a closeup of Michael C. Hall shadowed in blue and magenta light. 

My widget for “Bodyguard” was an angsty looking Richard Madden, and this account’s “Bodyguard” is an angsty looking Richard Madden in black and white. It’s interesting that the content of the photos are virtually the same but that the color schemes are not — perhaps even the lighting and dark contrast in photos is a purposeful appeal to masculinity. 

Then I see the widget for “Glow.” When I was recommended the show, it pictured two women: Alison Brie and Betty Gilpin in 1980s workout garb and neon lighting. On this man’s account, it’s two men (Marc Maron and someone I don’t recognize) looking annoyed on a tennis court. 

Men, I think. Of course, they’d only care about male characters. But just as I’m about to celebrate my feminist superiority, I notice that the man’s widget for “How to Get Away with Murder” has a blonde woman, while mine had a bearded man. If I were to ascribe a purpose to this difference, I would say that Netflix is trying to draw in the average heterosexual male with an attractive woman … but it’s hard to tell.                

Conclusion: There are a lot of TV shows about angsty white men with questionable morals. Besides that, the recommendations on the middle-aged man’s account were pretty legit, although the algorithm for which picture is displayed on the widgets still eludes me. 

While I began this expedition with the intention of showing how Netflix keeps people in bubbles, I no longer entirely believe this is true. A lot of my TV show recommendations are the same as this man. Perhaps the biggest difference between our Netflix accounts is how they advertise TV shows. If anything, given that customized advertising can draw a more diverse array of viewers to the same show, this is an argument against bubbles.

Either that, or I’m more similar to a middle-aged white man than I thought. 

Gabriella Del Greco SC ’21 is one of TSL’s TV columnists. She majors in Economics, and in her free time, can be found doing homework at the Motley and probably watching more television than is medically recommended.

Facebook Comments