This post is co-authored with Justin D Burton.

File under “not at all surprising”: we are pretty sure Beats Music is sniffing users’ Gmail and feeding that info into their “Just For You” recommendations. A few weeks ago I (Robin) mentioned to Justin that I thought this was happening to me, and then he discovered that it’s likely happened to him, too.

Justin: I found a most welcome message in my inbox a few days ago. I teach popular music at Rider University, and a friend who knows 1). winter break is hyper-writing time and 2). I’m always on the lookout for writing and thinking music (you know, a friend) was recommending a recent Hot Since 82 mix I might try. I wrote back to coolly express my gratitude (“OMG! Thanks so much for this!!!”), made a mental note to download it when I was back from traveling, and went back to thinking of snarky things to say about year-end music lists. A few days later, as I scrolled through my Beats “Just For You” section, hoping to find the perfect soundtrack for my morning coffee (*not* The-Dream), there was Hot Since 82’s 2014 album, Knee Deep in Sound. I’ve only been using Beats for a couple of months, so my “Just For You” list is culled from my listening habits in recent weeks (mostly Nicki Minaj, Azealia Banks, and Rihanna…okay, fine, also Drake) and music I may or may not much like but play in the classroom to critique with my students (this is how The-Dream and most of my rock recommendations find their way to being “just for me”). In other words, I’m very interested in Hot Since 82, but it’s not likely Beats would know that yet. Unless, of course, Beats had peeped my email. I like that Hot Since 82 is part of my profile now–Data Claus stuffed some much-appreciated variety in my JFY stocking. But I have this feeling that maybe Beats reading my Gmail isn’t always going to work out so well…

Robin: I made the mistake of hate-watching an episode of Dave Grohl’s HBO series “Foo Fighters: Sonic Highways,” and then I made the further mistake of emailing someone about how awful it was, the Rock Dads and the nationalism and the interview with President Obama, like he was a musicology Ph.D. not a J.D. I mentioned Dave Grohl by name in the email, and then within a few days, the Foo Fighters–who I have never listened to on Beats (nor have I listened to Nirvana…L7 or Ministry are about as close as I get)–were all over my “Just For You” recommendations. It felt almost like the algorithmic version of the U2 album appearing in my iTunes: here is some music that I really, really don’t want in my digital space, crowding up my music feed. (I mean, it’s probably not coincidental that Apple is behind both the U2 album and Beats.) It also felt a bit like a Rickroll: I was surprised with unwanted music where I least expected it to show up. (In retrospect, the old internet meme of Rickrolling seems like it foreshadows late 2014’s series of unwanted media objects dropped in users’ feeds or libraries.) More importantly, Beats’ recommendation algorithm seems to be weighing my emails more heavily than my actual behavior in the app itself (my favoriting, my searches, what I actually listen to)–but I’d need to know more about it before saying anything more definitive.

Like most uses of big data methods, Beats’ data-mining can make life more convenient (in Justin’s case) or more of a pain (in Robin’s case). Beats’ Gmail algorithm seems unable to distinguish between positive and negative mentions of artists: it couldn’t tell the difference between Justin’s praise and Robin’s disdain. It treated every mention as an endorsement or “like.” The Beats bot’s confusion about our actual preferences isn’t too surprising in a social mediascape that builds data profiles out of primarily positive affect. Our affective interactions on Facebook, for instance, are notoriously governed by the Like button. Reading together Sarah’s and Jenny’s recent posts about Facebook’s Year in Review videos, we’re confronted with a platform that has mined us to the point that it assumes the right to tell us a story about our own lives. And that story resonates through a filter of compulsory happiness, a narrow band of thumbs up that are meant to deflect critique, anger, grief, WTFs. When the Beats algorithm populates our JFY lists with artists referenced in our Gmail correspondence, it seems to be calibrated according to the same assumption: mention = like. Instead of compulsory happiness, Beats filters our chatter through compulsory fandom.

But, um, WHY? Why both assume and enforce compulsory fandom? There has to be some good business reason for this, right?

We’re used to social media collecting data bits about us for circulation in the marketplace. But Beats is ad-free, a subscription-only service. If Beats already has our money and doesn’t advertise other products to us, what does it gain by listening to our email conversations that it can’t gain from our listening and liking habits inside the app? And why does it hear these conversations as compulsory fandom?

It may just be that, given the limits and affordances of most big data business, compulsory fandom is the only way Beats can hear these conversations. It may be that there’s no good, reliable, or cost-effective way for Beats to hear mentions of an artist as anything other than an endorsement.

It may be that we’ve become so accustomed to algorithms tailoring streams–of social media posts, of advertisements, of “users also bought”–to our idiosyncratic tastes (or finely-tuned psychographic categories) that we expect Beats to adapt to us. Perhaps Gmail is just another source of input for Beats to use to make its tailoring more accurate than its competitors’?

Emailing my friend or coworker about an artist might count as a “recommendation within a personal network,” which is generally thought to be “far stronger than a non-personal recommendation,” such as one made by a bot or by a casual acquaintance. If so, then perhaps Gmail can be thought of as a hyper-personal social media platform, one where the “recommendation” found in a message would be read as the particularly substantive “wisdom of a friend,” somebody who you bothered to email and not just hit up on FB Messenger, on Twitter, or via text.

So, even though Beats is (probably) paying Google for all this Gmail data, it’s a sound investment for them if it helps them stand out from the competition and offer recommendations that really, really do appear “Just For You.” Though sometimes it ends up feeling like a specially-designed troll rather than an uncannily helpful suggestion. But perhaps that’s an acceptable risk for them? Maybe annoying some people some of the time is actually just the cost of delivering what Beats thinks is a better, higher-quality service? Perhaps, for users, compulsory fandom is the cost of having a more custom-tailored stream? Now that we know that Beats is listening back, we’ll start writing about music we don’t want to appear in our Beats streams using some of the same techniques that social media users employ to stay out of search results: next time, it’s “this F00 F1ght3rs series is barf” or “ugh, this Th3-Dr34m song.” (There’s an “Ima be fresh as hell if the Beats listenin’” joke in here…) In order to curate a pleasing stream of music on Beats, our email habits have to change. And that’s the other side of algorithmic culture: not only do they adapt themselves to us, but we adapt ourselves to them.