How Spotify, Netflix and Amazon control your online habits
18th Jan 2014 | 12:01
When is a choice not really a choice?
With thousands of TV and movie options and millions of songs on tap, it may feel like there is real choice. But these companies with their vast oceans of content have crafted powerful recommendation engines, taking heed of what we already like and drawing conclusions about what else we would enjoy. We gobble up the recommendations because they're so convenient.
But here's the thing. Netflix says 75 per cent of all viewer choices come from its Suggestions engine. That's a startling figure for a tool that has its moments, but often that kicks up more crap than the meanest bull at the Texas State Rodeo.
Better, fewer options?
Increasingly, on-demand content providers and web portals are using big data to tell us what to watch or listen to, or at the very least narrow down the options, supposedly for our own good.
Take Spotify, for example. Its entire platform is underpinned by a finely tuned trifecta of human editors, computer algorithms and social media data, bringing recommendations based on previous listening habits.
It's damn good at it too. While that's initially gratifying, there's a risk it could reduce the notion of musical free will to a mere illusion, if we come to rely on tools like Discover.
Donovan Sung, the project manager overseeing Discover and the wider recommendations engines that pervade across the service, told TechRadar: "There's a fine balance between providing too much choice or too little choice in this information overload society.
"Users don't want 100 options; they would rather have one great option or four great options. We err on the side of providing too few."
Where's all the new stuff?
Many have complained that Spotify Discover, which now acts as the service's homepage, doesn't do enough to furnish people with new music.
Rather, it takes the 'you like X so you'll like Y' which promises more of the same, and the 'you haven't listened to Y in a while, give it another spin,' approach. Again, Sung says user listening habits have informed these decisions.
"We've actually found that a lot of music discovery sessions aren't about listening to new music, they're actually about familiar music or things you've played recently," he said.
"Music discovery, in the case of discovering new things, is not a super main use case for most users, those are the power users who want to hear every indie band they haven't discovered yet. There are ways for them to do that beyond Discover."
Those users can use the App Finder to find curators dedicated to new music, they can follow influential playlist creators and there's even a new playlist the company is testing which draws together the most interesting stuff from everyone in the user's network. The Browse selection has a massive new releases section also.
So it's not that the new music isn't there, but us users are happiest when hunting within our own pack, according to Spotify. If we want new music, the onus is on us to seek it out.
'You know me too well, mate!'
Perhaps the problem isn't with bad recommendations because Spotify's are well thought out. In this case perhaps the suggestions have gotten so good we've stopped looking beyond them and, in terms of broadening our musical horizons, that's a problem.
So what's Netflix's excuse? It can't make the same claims as its musical equivalent when it comes to archive content (which it cares less and less about) or new content. It lost the rights to 85 classic movies at the start of 2014, that New Releases section has been showing Louie Season 3 for about 6 months now, and our "Suggested for you" barely changes.
A cynical mind might suggest Netflix is so keen on making sure we see its recommendations because there's not that much else beneath the surface of those "Because of your interest in…" match-ups.
Without the recommendations in place, you might use Search on the off chance you want to find the film you really want to watch. Netflix really doesn't want you to do that.
"Almost everything we do is a recommendation," said Xavier Amatriain, the company's engineering director, in a recent interview with Mashable.
"I was at eBay last week, and they told me that 90 percent of what people buy there comes from search. We're the opposite. Recommendation is huge, and our search feature is what people do when we're not able to show them what to watch."
Social status quo
But, it's not just music and video streaming saving us brain power with their well-tailored choice. Social networking is another great example of how curating so much content leads to us seeing the same thing over and over.
If you're using Facebook or Twitter for news and entertainment consumption, chances are you're following links posted or liked by your friends and/or people who share the same belief system, work in the same profession, support the same football team and so on.
It's basically like being in the same pub with the same people every Friday night. The conversation is different, but somehow the same.
Amazon's book recommendations were as organic as they come. User curated reviews and star ratings. Then in May 2013 it bought the reading-themed social network Goodreads. Kindle users can now see what their friends are reading and get personalised recommendations based on their previous tastes.
In loving memory of Myspace
There's a flip side to this 'choice through narrowed choice' debate though beyond the services and consumer. What about the content creators? Can new stuff stand out in this environment?
Kevin Douch owns Big Scary Monsters records, one of the UK's most influential indie record labels. He sees both positives and negatives, but said streaming services, and all that comes with them, have been beneficial for his artists.
"For new bands, the exposure can be great," he told TechRadar. "Being included on playlists, even just the perception that they're now a 'real band' when their music first appears in the catalogue, can be really helpful.
"Some of our bands, for example Gnarwolves, who have a very passionate, young and quickly expanding online fanbase, see their music streamed and shared a lot, and I struggle to see the downsides of having extra exposure when you're trying to establish yourself in what's a very overcrowded marketplace."
Of course, online music discovery tools aren't new. For the efforts that the likes of Spotify, Rdio, We The Hunted (which specialised in unearthing new music before it was bought out by…), Twitter Music and the soon-to-launch Beats Music have put into creating the perfect music discovery tool, nothing beats the original…
"I was saying just last week that we currently don't have anything as good as Myspace was at its peak, before the spam robots set up camp," said BSM's Kevin Douch.
"The simplicity of having your streaming music, list of tour dates, a brief biography and the top 8 friends (in itself a very powerful music discovery tool) on one page was brilliant. The addition of being able to comment and discuss via these pages made it even better."
"I think we need a new Myspace. By that I don't mean Myspace, I mean something that can get the mix right like they did all those years ago. The social interaction of Facebook, the fast pace of Twitter and the streaming capabilities of Spotify, but with the over-riding control left in human hands."
The Human Element
Spotify is careful to keep the human element in the loop, alongside its powerful data tools, and has a large editorial team feeding into its recommendations algorithms. Streaming app Slacker Radio says 84 per cent of its paying customers spend more time listening to human curated playlists than those generated by algorithms based on a song or artist.
However, if you think recommendations engines and discovery tools are controlling our listening habits now, streaming companies are only just getting started. At the moment, Spotify requires you to at least state some kind of intent. Pretty soon, it may not have to.
"The more we know about you, the better the engine can be," says Spotify's Donovan Sung.
"Maybe with motion sensors in phones, we can start guessing things like 'are you running, biking or driving?' Maybe it has a temperature sensor, or a heart rate sensor so we can get a sense of whether you're tense.
"Maybe it connects to some other services, for example if we know more about your sleeping habits we know what time you're likely to go to sleep or what time you wake up it can be personalised."
Enjoy the choice you have now because pretty soon these services might know you so well that you'll open the app and hear exactly what you want. And guess what? You won't complain.