is the Verge’s weekend editor. He has over 18 years of experience, including 10 years as managing editor at Engadget.
Posts from this author will be added to your daily email digest and your homepage feed.
This is The Stepback, a weekly newsletter breaking down one essential story from the tech world. For more on how to break free of the algorithm and discover things you love, follow Terrence O’Brien. The Stepback arrives in our subscribers’ inboxes at 8AM ET. Opt in for The Stepback here.
How it started
I used to have this ritual. Every Tuesday, I would get off the train at 8th Street on my way home from work. I would pop into Other Music, buy a new CD (or three…) and then walk the rest of the way to the Staten Island Ferry listening to my new CD. Even if there wasn’t a new record out that week that I was looking forward to, I would buy something. Often, if I went in without something specific in mind, I would consult the pyramid-shaped shelf at the end of the aisle that housed the staff picks. I would read through the index cards taped to the shelf containing handwritten endorsements from someone who worked there and grab something that sounded interesting.
It might seem like the ancient past, but until the 2010s, most people discovered new music in similar ways: browsing in a record store, friends at school, someone’s cool older sibling, CMJ New Music Monthly mix CDs.
That started to change in the 2000s with the advent of the first algorithmic recommendation engines. Pandora was the big pioneer there with its Music Genome project. The aim was to break songs down to easily quantifiable traits such as “gender of lead vocalist, level of distortion on the electric guitar, type of background vocals,” and the like. Then it would look for other songs that shared a certain number of traits in common and play that track.
Pandora enjoyed early success because its algorithmic approach to music recommendations was novel at the time. But there were also warning signs of problems to come. Anyone who used Pandora in the mid-to-late ’aughts will no doubt be familiar with its tendency to replay the same 10 or so songs over and over again.
This was in part due to it being at the forefront of streaming while having a small library. When Pandora filed its IPO in February 2011, it had just about 800,000 songs from 80,000 artists. Compare that to today, when even smaller players like Qobuz have over 100 million tracks.
Just a few months later, in July 2011, Spotify landed in the US with a catalog of 15 million songs and changed everything. Pretty much from moment one, Spotify was all in on algorithms. In 2015, Spotify launched perhaps its most iconic feature, the Discover Weekly playlist, which served up fresh algorithmic recommendations every week, as the name suggested.
... continue reading