My complaints make Netflix vulnerable

Everyone loves Netflix. From Wall St. to the bus stop on Wall St. everyone is watching Netflix. I’m a subscriber and in fact loved the service so much when I got my first envelope in the early 2000’s that I immediately bought stock. At about $15. I almost can’t believe that’s true – that it was ever that cheap. If you are one of those people who says “Look at their share price today, they can do no wrong” please stop reading.

The media has created a narrative that Netflix is an unstoppable juggernaut. Go have a listen to shows like Business Wars to hear how people are worshipping at the altar of Netflix. Every move they make is lauded as genius.

But I don’t believe they’re invincible. As a customer, I’ve observed some moves they made and see they are trying to obscure their weaknesses, which are really quite obvious.

Recommendations and data

Netflix has become legendary for data. Not just legendary, but Barney Stinson “Legen – Dary.” It is clear that they collect a lot of data, what we watch, for how long, how many times. How long we are exposed to the opportunity to view a show before we save or watch it.

The most oft cited example of their data prowess comes from House of Cards. They used troves and troves of data to get to this amazing conclusion. Are you ready? I don’t think so – it’s huge. A highly awarded and regarded hit television mini-series from the UK would be successful as a US translation. To further push their data-based conclusion, they uncovered the shocking idea that it would be more successful if they hired one of the most influential and visionary directors in the world (David Fincher) and a name brand, A-list actor to headline the production (Kevin Spacey).

Can you imagine the data they had to sift through to get to this amazing combination? In fact, Universal got to the same conclusion in 2009 without the benefit of all of Netflix’ data when they distributed State of Play. A film based on an award-winning BBC series, with an A-list actor (Ben Affleck) and a strong director (Kevin Macdonald, admittedly, he’s no Fincher but you get the point).

I’m sure they are using their data in ways that users don’t see. But many of the ways that I do see are awful when held up to the legendary status. If you are a Netflix subscriber, go look at Netflix top picks for you. Mine have very little relationship to things I have watched and rated well. Or at least if there is a relationship, I don’t understand it. There are obvious things like the same director or actor in something I watched or gave a thumbs up to.

Sometimes, it’s a similar category. I watch horror, so it makes sense to recommend horror. If I watch something from the early 2000’s like L.A. Confidential, I understand why they would recommend other genre content from that same time or other noir.

But I haven’t ever watched a romantic comedy, so I don’t know why it would ever recommend one. Every now and then I watch comedy, so Netflix punishes me relentlessly with stand-up recommendations. They seem absolutely random and almost desperate. You like this one, right? RIGHT?  It’s not based on the comedian or the style of comedy. And there’s never a connection that’s shared visibly. I don’t know what stitches the recommendation to something I’ve watched and liked.

Netflix frequently recommends content I’ve already watched on Netflix. Not the “Watch It Again” variety. But just trying to sneak it by. Maybe he won’t remember. I also get rows and rows of crazy recommendations in the “Because You Watched Movie X” that have zero to do with the content being referenced. See below:

Yup, this is a partial list of all that were recommended based on Goodfellas


Netflix’s library is vast. There are movies and shows from hundreds of countries. With a count that high, it’s odd how frequently users can’t find something to watch. Though on-demand programming is an order of magnitude better for viewers than the timetable-based programming of cable, it’s created a different problem.

We now believe we always have something we demand. We used to turn on the television and flip channels mindlessly until we landed on something we liked enough to put down the remote. Now we believe we have our every viewing desire at our fingertips. But instead of flipping channels, I find myself descending down rows of loosely grouped options until I find something I like enough to put down the Roku remote.

When someone tells me a film was nominated for an Oscar, all I hear is “it isn’t available on Netflix.” Part of the allure of the brand is that there is an essentially limitless amount of film and video to watch. Except the really good stuff. In 2002, I used the service to get familiar with all (or most) of Takashi Miike’s and Park Chan-Wook’s catalogues.

Now it seems really hard to find something to watch. There’s something wrong because there has never been as much available on the site than there is today. But it falls into the paradox of choice. The more options, the more users become paralyzed. Watching Netflix through Roku, you can scroll forever before you find something to settle on. This is at least in part because users are concerned that they may miss the best choice.

But logging on to the service with something in mind is even more frustrating because searching the catalogue absolutely sucks. The best bet is to search on mobile or your laptop until you find the exact thing, adding it to your list and then going back to TV to pull it up. It’s not like any of their competitors in video content have highly integrated search features, so I’m sure I’m overreacting here.

From ratings to matches

Up until mid-2017, every piece of content on Netflix had a star rating. Users understand star ratings. 5 is good. 1 is bad. If I’m taking a chance on something I’ve never heard of, 4 or 5 stars seems to be less of a chance. When Netflix made the move to swap star ratings for match percentages, I immediately took notice. I’m pretty clear on what the match percentage is; based on my viewing history, I assume the match percentage ranks how likely this is that I would select this thing. I’m also guessing it’s based on common content viewed by others who watched things I’ve watched.

OK. A 98% Match seems like it will be a home run. But what is a 76% Match? Is that good or bad? Let me look at the star rating to decide, oops it’s gone. Without that context, how can I judge? Sometimes you can find a reference to the trigger content: Because You Watched X. Because I Watched Kill Bill Volume 1 Netflix recommends Kill Bill V2, Inglorious Basterds, Sin City, Donnie Darko, Scarface and Casino. Makes sense. Good use of data. In the same group of recommendations, Netflix is also offering Parks & Rec, Tomb Raider, Disney’s Hercules, The Office (US), Queer Eye and Gossip Girl. Not so good use of data.

If I spend a minute putting thumbs down ratings on things I’m not into, I’d expect they’d be replaced with new options, or at least removed. But they aren’t. They remain offered, although they are sometimes grayed out during the next session.

The reason for the switch is actually to confuse me in the exact way I’m confused. When I was looking at star ratings, I would quickly rule out a lot of stuff under 3.5 stars. Now, I’m seeing options and I don’t know exactly why but there are still a ton of options and I have to think a bit more about ruling them out. In the Star system, I started to get the sense that there weren’t that many 4 or 5 star options available on Netflix. In the match percentage model, Netflix is saying “But look at all this stuff that’s just right for you.” 

Which leads me to this final point.

Most of their original content is not great

Netflix is praised by content creators for their hands-off attitude. They bring in auteurs to execute their vision and let them go. Definitely worthy of praise, and should be continued with some caution. For every Stranger Things there are 20 The OAs. Things that sound great as a synopsis but aren’t really worth watching.

Digital video production has leveled the playing field. It’s relatively simple to get great looking footage without having to deal with all the expense of film. But storytelling isn’t that simple. A lot, read: a fucking lot, of the Netflix original content is just unwatchable. That’s my singular opinion. Except it isn’t. As a whole, the content is average at best.

Bad content isn’t just a problem, it’s the big problem that’s creating all of these other problems. Netflix is forcing content creation to build separation from Disney, HBO et al. When they moved to streaming, the Starz catalogue may have been enough. But now they have to protect their subscriber lead from the giants – who kill it at telling stories beautifully. They’re sloppily adding content they own outright.

To shield themselves, they wanted to distract users from an onslaught of unwatchable content. They rid the system of star ratings that were easily understood and swapped in match percentages. This was done to hide the scores and play up their legendary data strength. It was also done so people would attempt to watch this shit content.

They could certainly improve the search function to make wading through piles of content faster. But they want users to believe there is so much to choose from more than they want them considering just what it’s so much of.  How could I quit Netflix, there’s so much here. It doesn’t matter what Disney adds (all of its own legendary content, new exclusives PLUS all of Fox content including FX series that they have been very cagey in hiding from free streaming).

Are they doomed?

F-ck, no. They’re in a strong position. These are nitpicks.The experience and the complaints listed above have started to make Netflix feel a lot more Netflix-centric than user-centric. One way they might add to their existing advantage is to push UX. Improve search to make it easy to actually select something and get watching instead of wading through endless rows of options. Lord, please kill the autoplay previews upon initial login. When I select a show from the list and it plays, I’m good with that. Keep the new show promo assault toned back just a little.