Stop me if this sounds familiar: You have the TV on, the clicker in hand, the popcorn made, the lights dimmed, and you're ready to watch a movie. You fire up Netflix, and start scrolling. And scrolling. And scrolling. Suddenly you've blown half of your two-hour window just looking for something to watch, drowning in a sea of star ratings and Western Movies With Strong Female Protagonists. Hate to break it to you, but you've come down with a case of Netflix Paralysis. It's an epidemic.
A new app called MightyTV (iOS-only for now, Android soon) aims to solve this problem using a unique concoction of Tinder-style swiping, machine learning, and a collaborative personalization mechanism that can not only find something for you to watch right now, but can account for the tastes of everyone on the couch and find something you'll all love. The app is out today, it's super slick, and it might have a chance to do what so many have tried and failed to pull off.
MightyTV is the invention of a team led by Brian Adams, an entrepreneur who spent much of his career working in ads. He sold his company, AdMeld, to Google in 2011, and then spent the next few years working on Google's money-printing Doubleclick ad system. In February of 2015, he quit his job and started poking around things like local search before landing on video. "The problem I wanted to solve was discovery," Adams tells me, leaning back in his chair with a phone in his hand running a customized-for-him version of the Mighty app. "I felt it was getting a lot more difficult."
He's not wrong---Jim Nail, a principal analyst at Forrester, agrees. "I think the next thing that people need is some way to even find and corral all of the content that is available to them out there," he says. Not just with video, either; with everything.
The MightyTV app centers on a few core features. The first is essentially Tinder for video: You open the app, and see a large poster for a movie or TV show. Swipe right if you like it, left if you don't, up to skip, and right and hold if you love it. This is how Mighty figures out what you like, which is both the most important and hardest part of building a tool like this. And there's just something magical about the Tinder interface, which just makes you want to swipe. "People usually do at least, like, 100 swipes when they first start," Adams says. Mighty gathers listings from your subscriptions (Netflix, Hulu, HBO, and others) and it pulls data from Facebook to get some knowledge about what you like. It does some sorting based on rating and release date, but it mostly relies on your own swipes. There's just no substitute for user input when it comes to movies. A 2009 study, prompted by the infamous $1 million prize Netflix offered for a 10 percent improvement in its recommendation algorithms, found that "even 10 ratings of a new movie are more valuable than the best solely metadata-based representation." People like movies, the authors wrote, not their descriptions or genres. And when you know what a lot of people like, you can pretty reliably start to guess what others might.
The backend of MightyTV is powered by a series of machine-learning algorithms that take your swipes and try to learn your taste by comparing you to the rest of the service's users. As you give the app more data about your tastes, and your friends do too, it starts to figure out how different content relates. Why they relate is a different, harder question, one Adams sidesteps for now. "What I'm working on is a method to figure it out without really knowing," he says. "It's like acupuncture: I don't need to know why you like something, but if I can predict that you will based on 100 other people you don't even know, then that's going to improve your life." Essentially, Mighty's algorithms use volumes of data---what users like, what they put on their watch lists---to make connections between things. Over time, in aggregate, the software can start to make guesses. If 100 people like Casino Royale and 99 of them also like The Bourne Identity, you can fairly safely guess the last person's going to enjoy it too. And if they don't, or they've seen it, they can just swipe left. Tinder's not right every time either, after all.
Right now, a feature called Mashup offers the most immediate hint at the technological sophistication inside Mighty. If you're sitting on the couch with your boyfriend and best friend, and you're all Mighty users, you can open up the app and tell it you're all looking for something to watch together. Mighty then combs through what it knows about each of you, and recommends a movie or show it thinks you'll all love. When people asked Adams why he was starting the company, "I used to say, I want to find the best thing to watch in a minute or less every time. And I want to be able to do it with my wife too." I've been testing out a beta version of the app, so none of my friends are also MightyTV users, but Mashup could be big.
Finding people something to watch is a noble goal. So noble that basically everyone is trying to do so. There are plenty of streaming guide apps, like Yahoo Video Guide and Fan.TV. There was QPlay, from the founders of TiVo, that tried to mix cable and the Internet. When I spoke with Hulu head of experience Ben Smith last fall, he said the service is spending a lot of time trying to turn long lists into actual recommendations. Their goal is blissful brainlessness for users: "that first two or three things that you see, you just say that's what I want to watch. And then we get back to rewarding that inherent awesome laziness of TV." Right now, sorting through everything is an obvious pain point for so many people, especially compared to the mindless channel-surfing we're used to. "It's just too much work to think about what am I in the mood for tonight," Forrester's Nail says, "and spend ten minutes going through the list and searching for something."
Most streaming guide apps are simple search engines. MightyTV is that, certainly: it can find you things to watch across Hulu, Netflix, HBO, and others, and take you to them with one click. It also has trailers for everything right within the app, which is wonderful. You can keep a list of things to watch, and filter everything so you only see TV shows that, say, cost less than two bucks. That's table-stakes for an app like this, Adams says. Where it gets interesting is when your guide can be predictive. Not just by showing you lists of popular movies, or list of movies starring actors you seem to like or with plotlines similar to that one horror flick you watched six months ago. By actually understanding what you like, and telling you what you'll like next. If Mighty can solve that, it'll take binge-watching to a whole new level.
