What’s the Yams?

I’ve always loved music. From a very young age, creating and listening to music has been an integral part of my life, and it remains so. So, you’ll understand if this is a bit long-winded — music matters a great deal to me.

I’ve carefully built an iTunes library from the time that I bought my first iPod in 2005 with music discovered in lots of different ways, from automated services like Last.fm to personal recommendations from friends and family.

One of the services I’ve long used is Apple’s Genius feature, built into iTunes since 2008. Since its introduction, I’ve provided Apple with the entire contents of my library on a weekly basis.

So it will come as no surprise to you that I have not been entirely enamoured with the recommendations provided in the “For You” section of Apple Music. For the uninitiated, Apple Music functions entirely separately from its iTunes precursor. It uses a different library of music — though they’re almost identical — with a separate cloud storage mechanism. Even its recommendation engine is detached from iTunes, which means the past seven years of data about my library is completely forgotten.

And that, in a nut, explains why I keep seeing “Intro To…” playlists for Nine Inch Nails, Kendrick Lamar, Kanye West, Modest Mouse, Burial, Deftones, and other artists for which I have every record they’ve made. iTunes, by way of Genius, knows this. While I have listened to every song from those records at least once, iTunes knows which songs I skip regularly and which songs I’ll listen to time and time again. But Genius doesn’t talk to Apple Music, which is why my recommendations are inconsistent and frustrating, despite tapping the little heart icon on everything I enjoy.

Other algorithm-based suggestion services have also found me wanting. Perhaps it’s because computers interpret patterns differently than you or I might, making it difficult for them to be programmed to suggest music we might like, but not necessarily in a way that’s directly related to something we already enjoy.

So I’ve been hunting for a new way to find new music, but it can be hard. Everyone has different tastes and approaches new music with different expectations. You may have a multifaceted aural palette, but you may also be exploring a genre or artist for the first time. Other people may prefer very specific sub-genres of music.

From the time I got my first iPod, I’ve been reading industry publications, scrobbling my played songs, jumped between a bunch of grey-area music blogs, and juggling a bunch of other ways of surfacing new music. But all of these things require a considerable time investment that I don’t necessarily have any more. I don’t want to have to train anything either. I just want to listen to great tunes.

There is not a universal solution, I don’t think, but I’ve found something that gets pretty close to perfection. It’s called the Yams. I’ve been using it for the past few months, I think it’s terrific, and I think you should know about it. And that worries me a little bit, for reasons I’ll get into.

I’ve said some critical things about algorithmic playlist builders, but I’ll add one more to the pile: they feel pretty anodyne. By contrast, personal recommendations from a friend or a knowledgable record store clerk feel, well, personal. These are people who know their shit, they know what you like, and they’re giving you an earnest opinion. And I love that.

Along the range of almost entirely algorithmic recommendation engines — iTunes Genius or Last.fm, for instance — to completely personalized — like that one friend of yours who always knows the best new stuff — the Yams feels entirely new because of how close it gets to the latter. The user experience is what defines this; the Yams exists entirely through text message conversations with a real person.

I was introduced to one of the service operators in a welcome text. His name is Miguel, he makes music as the Miserable Chillers, and he writes for Ad Hoc.¹ Good start. Here’s the first conversation I had:

What kind of music would you like to hear?

I have a varied taste in music, though I typically stay out of the top 40. I’d like to listen to more interesting rock (of all genres, with the exception of chugging post-grunge and anything much louder than Deftones or Mastodon), and more interesting hip-hop and rap (I love Kendrick, Kanye, RTJ, and Odd Future). Plus, anything experimental and weird you can think of.

Cool. We can definitely dish out some weird stuff. Do you use any streaming services?

Apple Music and Spotify.

I kind of lied here — I have a Spotify account, but I haven’t used it since Apple Music launched, and I discontinued my premium subscription as soon as I received my first bill from Apple.

Perfect. We’ll get back to you soon, Nick.

A little over half an hour later, I was sent a link to a personalized Spotify playlist: “Y.O.D. – N. Heer”.² As it was an introductory playlist and I provided a wide swathe of genres and artists, I received ten tracks that span the gamut. A mix of bluesy rock, electronic bleep-bloop stuff, some great indie hip-hop, and a little fuzzy psychedelia played for a little over forty minutes, and it was wonderful.

Every so often since, I’ve sent Miguel a text with a request for a new playlist. Last month, I was working a little late and I asked for a playlist based on Burial. A couple hours later, I received nine tracks that were completely in the same vein. Yes, there were some gimmes in there — Four Tet and Zomby — but they also popped in some XXXY and FaltyDL, both of which I’d never heard of before and thoroughly enjoyed.

Then, a couple of weeks ago, I wanted a playlist to match the weather, based on my previous requests. I got back a truly excellent melancholic playlist that was blessedly free of Bon Iver and Death Cab. Don’t get me wrong — I dig both of those artists, but I’ve heard enough “rainy day” playlists that feature them.

I think that gets to the crux what I’ve enjoyed most about the recommendations from the Yams. On other services, especially more algorithmic ones, songs and artists that are popular will tend to be recommended more often, thus becoming even more popular. The playlists I’ve received from the Yams, on the other hand, have been chock full of interesting, off-the-wall tracks that I haven’t heard before. It’s truly music discovery.

The other thing that makes it so great is the user experience. The conversational aspect of it is what helps it feel so much like asking for new music suggestions from a friend.

So why am I worried about the Yams? I have a hard time seeing how the text messaging and personalized service will scale to work with a very large user base. I want to tell everyone about it because it is truly great and innovative, but I worry that this thing that I use and love will become saturated and lose that personalized quality. In some ways, that’s already happening a little — I have occasionally felt the need to text a reminder when a playlist I requested was seemingly forgotten about.

I wanted to know more about this, so I asked Miguel who I could talk to. He suggested I email the founder and CEO of the Yams, Shannon Connolly; yesterday, I had the opportunity to speak with her by phone.

She told me that they’re structured to deal with scale. Front-line operators like Miguel, Connor Hanwick, and herself dispatch playlist requests to contributors who know what they’re talking about. Every contributor has a generally great level of music knowledge, but they all have their own areas of specialty: some are into funk and soul, while others are metalheads.

Not only does this help with scalability, it also allows for a future personalized recommendation service for pretty much anything. Want to know what book to read next or movie to watch? At some point the future, you’ll be able to text them about those kinds of media as well.

One of the things Shannon and I talked about at length was the advantage of the personalized user experience. While the Yams uses bots to automatically reply for generic messages — such as when a request is made outside of business hours — all of the interactions that occur while making a request are with an actual human being. That means that your time is not spent fighting with a bot.

The humanity of the service also creates a deeper level of interaction. Shannon told me that they’ve had users text them about specific moments in their life, like breakups, getting married, or being fired from their job. I can see another advantage to this if you want a playlist that specifically omits some songs or artists.

Aside from my concerns about scalability, the only downside I can think of is that it’s not an always-available service, as I alluded to above. It relies upon human operators, so it only really works within business hours, and it does take minutes-to-hours for a playlist to be created. So if it’s 11 PM on a Sunday, there’s little chance you’re getting that emergency study session playlist. (You can add “URGENT” to a request, but there’s no guarantee.) Their organizational structure allows them the flexibility of adding more contributors and operators around the world, so this should be alleviated.

But I think that’s okay. The Yams doesn’t feel like it’s set up to be the sole place where you find music to listen to. It’s the thing you turn to when you need something new and different. Miguel is the guy who recommends new music to me when I ask. I’ve never met him, but he knows what I like to listen to. That’s pretty cool.

The Yams is currently in a private beta, but they’re adding new users all the time. You can register on their website.