Late Night Dispatch: Tuca & Bertie versus The Algorithm

Jul 29 · 10 min read

[A quick note!! Hi! I used to write a newsletter, but I stopped because life got busy. Lately, I haven’t been able to sleep very much. When I can’t sleep, my mind obsesses over “hyperfixations,” a term those in the ADHD community will recognize. So, I blog. I blog because it helps me work out these ideas, and tires me out. I’m going to use this new newsletter, “Late Night Dispatch,” as just that — writing through insomnia. This won’t be a nightly, weekly, or monthly thing. It’ll just be for when I can’t sleep. If you like reading my tweets (lol) or my work (thank you, if you do!!), now I have this. Also note that since none of this is edited and is written in the early hours of the morning, it’ll be incoherent at times. But that’s what late night dispatches are.]

Thursday, 9-10pm ET. That was the golden ticket to a successful show a decade ago.

There are hierarchies at networks. Anything that you could sell ads on, but no one particularly wanted to watch, went either before 7pm ET or after 1:30am ET Monday thru Thursday. Friday and Saturday nights were reserved for shows already on their way out the door. The only spot more coveted than 9-10pm ET on a weekday was 9pm ET on a Sunday. Maybe. Don’t take my word for it — look.

It’s still mostly true for the major networks. People who still tune in to television every week experience a curated watching experience, handed down to them by network programming executives, who study weekly ratings and people’s behaviors. That’s why a show like The Big Bang Theory or The Walking Dead receive guaranteed prime time slots every season.

But, really, what does prime time mean to us? I don’t own any form of basic cable. Neither do my friends. I can’t remember the last time I channel surfed. My version of coming home from work, flopping on the couch, and browsing is a three-step process: switch the PlayStation 4 on, select the Netflix (or Hulu or HBO Now or Amazon Prime Video) app, and rummage around for a couple of minutes.

I used to feel like this gave me more freedom. The choices were endless! There was too much choice! It became a pseudo-condition among cord cutters: a newfound form of decision paralysis. Eventually we’d all give up and watch Friends or The Office (similar to landing on Fresh Prince of Bel-Air re-runs on TBS) — or we’d pick whatever Netflix highlighted for us. These titles, a never ending barrage of movies and stand-up specials and TV series and documentaries — hand picked for us by Netflix’s impressive algorithm.

Finally, technology saved us from ourselves. And saved us from boredom. Netflix’s algorithm learned from each one of our choices, and promised to serve up something equally delicious for us to waste two hours watching. But it failed. It has been failing. It just took us until Tuca &Bertie’s cancellation to show us just how much it’s failed us.

Tuca and Bertie

I am the optimal subscriber for Netflix to recommend Tuca & Bertie to. Not only have I watched every episode of BoJack Horseman more times than I can remember, but I watch every cartoon Netflix releases. I also just watch a ton of sad shit. I’m trying to turn Netflix into my own personal Tumblr, but with less scrolling.

The algorithm never recommended Tuca & Bertie to me. I had to search for it. And I mean search. I basically took a James Cameron-style submarine deep into the Atlantic Ocean searching for any sign of Lisa Hanawalt’s animated series about two friends in their ’30s trying to get through day-to-day life. It should have appeared at the top of my carousel. Hanawalt worked on BoJack Horseman, the animation styles were similar, and it fit into the overall wheelhouse of what I watch.

Finally finding it, I marathoned the hell out of it, tweeted about how great of a show it was, and then went on with my day. It wasn’t until Hanawalt announced last week that Netflix decided not to renew the show for a second season, and this specific tweet, that I remembered my own difficulty finding it:

The cancellation wasn’t exactly shocking. Netflix executives have reportedly been telling middle-to-high-tier members of their team to be careful with spending. The internet network, which is still arguably the most successful companies in the streaming era, was amassing more debt every year. The loss of licensed series like The Office and Friends to rival streamers meant the network has to invest in more original content — and while that’s always been the plan, Netflix doesn’t have unlimited funds.

So, like a traditional TV network, Netflix does look to viewership numbers to try and figure out what shows are worth saving and what aren’t. Again, that’s not the issue. The company reportedly came up with a “two season” rule, according to The Information. If a show couldn’t pull in new subscribers or catch the attention of current subscribers after two seasons, the likelihood it could at all began to slide. Everything comes down to math.

The issue is that Netflix executives have to rely on an algorithm to make that decision in placement of an ongoing loss seen by shows in traditional prime time slots. If a show was front and center on a network, but never found an audience, that’s one thing. It’s unclear how many people were ever served Hanawalt’s show, and how many didn’t even know it existed.

This is essentially what Hanawalt and her team ran into. Despite people finding her show everyday through word of mouth, being called a critical darling, and even building a dedicated fan base, Tuca & Bertie was cancelled because not enough people saw it.

Netflix is still reliant on ratings, (collected internally as opposed to by a third-party company like Nielsen) but it’s unclear how those numbers are collected, what marks a successful audience turn out, and what’s a total failure. Even showrunners are kept in the dark. You have a show one day, and then one day you don’t.

The problem, as simplified as it is stripped down, is quite literally that not enough people even knew it existed because everyone assumed the algorithm would recommend it to the right people. It didn’t.

It’s a problem for everyone

Let’s go back to traditional TV for a second. Networks know pretty quickly if a show isn’t going to work, and they play their own game of hide-and-seek. A show that debuted on Tuesday at 8:30pm ET might be moved to Friday at 10pm ET to make room for another show that performed better than expected. Pilot season is a game of chess.

Some shows get cancelled halfway through the first season, but many get a four or five month period to try and find an audience. If it doesn’t work out, there are weekly ratings that people can point to. Because a show is on at the same time every week on the same network, for the most part, there is a definite visibility to it.

Tuca and Bertie was seemingly hidden from an audience who would devour it upon the first week of its release. In the entire time it was out, or leading up to the show, there were three tweets about it from Netflix. Hanawalt and her team hoped that Netflix’s algorithm, which promises a unique, tailored experience for every single person, meant their show ended up on the home page carousel of at least some people.

Did it? We still don’t know how many people were served it and who wasn’t.

Not every algorithm is the same, but there are similar core foundations to Netflix’s tool and what YouTube uses. They want to keep you engaged, and on the site, so they’re going to recommend two types of content: stuff that you’re interested in based on what your watch history and, perhaps more visible, what other people are watching. YouTube has gotten very good at the former. My YouTube recommendations are mostly spot on (albeit it with some outliers). Based on my own experience with Netflix, and the conversation that spurred in wake of Tuca & Bertie’s cancellation, Netflix’s algorithm is very good at the latter.

Kyle Chyaka said it best in his own newsletter on the subject, writing:

There is so much content on streaming / content platforms and so much more all the time that the services have the burden of delivering it to us in such a way that we don’t feel overwhelmed by it. There were a million channels on cable TV, but they were each only playing one thing at a time. Thus The Algorithm is completely necessary to navigating it all. It acts as a filter so we’re not totally blindly groping in the pile.

Using the carousel to highlight top series and new movies from creators like Adam Sandler — someone Netflix executives paid a lot of money to have on their platform — is the same as ABC using it’s 10pm slot on Wednesdays and Thursdays for its best shows, or HBO running Game of Thrones at 9pm on Sundays. There’s no issue with that method; it’s worked on traditional TV for years.

The problem is that Netflix’s recommendation algorithm is flawed — very. Again, based on people’s tweets over the last few days, people who wanted to watch the show had to seek it out. Other people didn’t even know it existed. This isn’t a problem that just plagues subscribers, but creators, many of whom reached out to Netflix on Twitter to air their own grievances with similar experiences.

At least with network television, there was a chance you could stumble upon something while channel surfing. It’s linear. Everything is there. There may be hundreds upon hundreds of options, but everything is lined up for everyone in the country at the same time. Netflix is like playing a game of hit or miss. Something could be there, or something couldn’t. I come across new series and movies every day that I just didn’t even know were on the platform.

Hell, I didn’t even know a new season of Daredevil came out until it was over and I saw an article breaking down the finale a week-and-a-half later. And I binged the first two seasons the first day they came out. Maybe Netflix tweeted about it, but tweets are ephemeral. Even trailers get buried on YouTube.

The thing that’s scary about Netflix — to subscribers and creators alike — is that if you’re not Adam Sandler or Stranger Things, there’s a good chance your show just won’t appear on Netflix’s home page. And if it’s not on the homepage, it’s as good as gone. No one is dumpster diving on Netflix, for lack of a better phrase, to find hidden treasures. Now that Netflix is operating like a network and reining shows in after two seasons, you only have 18 months to make it work.

Where do we go?

There is no future other than streaming. That’s not to insinuate that network or cable TV is just going to poof into nothingness. NBC is going to be just fine. But in terms of what matters, of where the industry is heading, of where our eyeballs and attention are, streaming is it.

No more prime time slots, no more public ratings, no more channel surfing. And while that may not seem like a big issue right now to us, the consumer public, it means that shows like Tuca & Bertie, from creators with voices that streaming was supposed to help, are going to sink. It’s a tough problem, one that I won’t pretend to know the answer to, but I know that we can’t rely on algorithms to tell us who’s watching shows and who’s not. If I’ve learned anything covering YouTube it’s that there are few things as flawed and dangerous as relying on an algorithm to figure everything out.

Here’s the future of streaming, as far as I can tell, based on a piece of anecdotal frustration. Tuca & Bertie was cancelled on the same day that trade publications reported Netflix was in talks to sign Game of Thrones creators David Benioff and D.B. Weiss to an overall first-look deal for more than $200 million, competing with Amazon and Disney to do so.

Benioff and Weiss are brilliant, but they don’t have any TV shows in the works right now (other than a possible show about the confederation at HBO that looks like it may never see the light of day). They’re signed on to make not one, not two, but three Star Wars movies for Disney. But because they made Game of Thrones, a show I deeply loved but whose last season was far from critically adored, they will possibly receive $200 million to make whatever they want for Netflix. I don’t know how much Hanawalt for Tuca & Bertie, but I feel safe assuming it was far, far, far less than that.

Most frustratingly, and the whole issue at the heart of this piece, is that if Benioff and Weiss join Netflix, you can guarantee whatever show they have will receive more than three tweets from the official Netflix account. It will receive more than a half-hearted attempt to promote a show.

You can abso-fucking-lutely bet that it will be at the top of every single Netflix subscribers home page for a full week when it comes out. That’s 150 guaranteed eyeballs. How many of those did Tuca & Bertie get? Only the algorithm really knows.

Julia Alexander

Written by

Nonsense thoughts about technology and, like, life, I guess.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade