Why Don’t Recommendations Look at the Bigger Picture?

Most people agree that the days of the mass cultural phenomenon are probably over. One in five American households are never going to watch the same thing, as they did with the Motown 25 concert. There will still be huge bands, but they’ll be huge within a particular crowd — teen pop bands, country superstars, etc. Every so often there will be a blockbuster film, like Avatar, but even those are rare — and compared to something like Titanic, which it seems like everyone saw, plenty of people skip the big hits every summer and come through just fine.

Instead, we are sectioning ourselves off into smaller and smaller niches, and everything we consume within those niches tends to align. Hipster Harriet likes prestige TV dramas, indie rock, mumblecore films, eating vegan or paleo, and dense novels that grace the cover of the Times book review. Redneck Rosie likes Duck Dynasty, country music, blockbuster comedies, fast food, and books that appear on the Times bestseller list.

Obviously, this is super reductive, and plenty of people like a big mix of things outside what they’d be expected to like based on income, location, and education. But take a look at yourself, and you’ll probably see that more often than not, your consumption patterns tend to align in a very particular way.

Which is why I’m honestly shocked that no one in the entertainment space seems to have figured this out. When I went on iTunes to buy a few TV shows to watch on the plane ride home from SXSW, the site was totally siloed into film, TV, and music sections — with no recommendations based on what I liked or purchased in other parts of the site. I’ll say upfront that I’m not a huge iTunes user, so the information could have been incomplete — but regardless, iTunes still has my credit card info and my address, and they could at least glean something from that. A Williamsburg zip code should be a pretty dead giveaway about my potential preferences.

It’s not just Apple — Amazon fails at this, too. Their recommendations within categories are usually solid, but they’ve never recommended something in a different category based on an order. I ordered the Kim Gordon memoir and it didn’t even take the obvious step of recommending Sonic Youth albums. I dug around and the only site I could find that offered cross-platform recommendations required you to enter them manually, and was woefully incomplete — when I entered Karl Ove Knausgaard, the site had “never heard of him;” entering TV on the Radio got me a whopping two book recommendations, both by well-known authors.

I connect my Spotify account to my Facebook page and I’ve never had them recommend anything based on any of the content they could possibly glean from it. And I’m not even talking about the deeper stuff they could pull — for instance, that I loved a recent episode of the The Americans that featured a Fleetwood Mac sync, and that it might be a good time to revisit “Rumours.” I’m talking about really simple things, like “you liked the Kurt Cobain documentary, how about listening to Nirvana?”

What I find galling is not that I have so much information about myself on the web — I put it all out there — but that no one can use it in any meaningful, smart way to serve me good content. My Facebook page says that I like doing crossfit, distance running, and eating clean; perhaps serve me with some great workout playlists (and don’t serve me ads for juice boxes, good lord). Facebook’s ad algorithms are clunky as hell, as well as being straight out of the 1950’s — as soon as I got engaged I got wedding ads, once I changed my status to married, here came the baby ads — and I’m surprised they can’t figure out something better.

Which leads me back to Amazon and Apple, and how they could use this as a real way to set themselves apart and win the streaming wars. As much as I like Netflix, they don’t have access to music; Spotify only has music, no movies or TV. But Amazon has everything, and they know everything. My doorman probably hates my guts because he signs for five Prime packages for me every day. Amazon knows where I live, what shoes I buy, and what my dog eats — and they can’t figure out how to put that together and recommend me some new music.

I do remember Target doing this with circulars a while back and catching flack for outing a pregnant teenager based on her purchase history, and maybe the recommendations I’m seeing are just so sneaky and under the radar that I’m missing them. I kind of doubt it, though — my guess is that companies just haven’t figured out how all the pieces fit when it comes to recommending culture. But there’s really no downside to giving it a try — no one will stop using Apple or Amazon if they recommend some books or films they don’t like. In fact, they’ll probably offer feedback that will help improve future recommendations.

And this is where Apple and iTunes and whatever Beats becomes can really kick ass. Apple doesn’t have as much as Amazon, because they don’t sell as much as Amazon, but they have enough in terms of credit card info and film, TV, and music consumption. I’d love to log in and see my whole cultural footprint reflected in the content I’m being served.

A lot of people I’ve interviewed have complained that streaming services are too focused on genre, when they really like multiple types of music that fit a particular mood or experience. Hopefully, by taking into account other things that people consume and learning how they all fit together, a company could provide recommendations that reflect a bigger picture.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.