Here’s the pitch:
If you turn on Spotify’s Discovery Mode for a song, that song will be placed more frequently in personalized playlists (like Radio, Autoplay, and Mixes). Hypothetically, this means more people will hear it.
Of course, there’s a catch.
First, your song has to qualify for the program (you can read the criteria here). More importantly, in exchange for the boost, you’ll give up 30% of your royalties on streams of the song from those contexts. But you’ll keep normal royalty rates when your song is streamed in other contexts.
If you’re thinking, “Huh, that all sounds a little confusing,” or maybe “Wait, isn’t that pretty much payola?” – then hey, welcome to the club!
(And in both cases, I’d say the club is full of people who are right.)
I’m writing today’s email because I get asked about Discovery Mode a lot, with most people wondering some variation of, “Well, is it worth it?”
My answer is that, for you, yeah, it probably is worth it. For everyone else, though, it probably isn’t.
Here’s what I mean.
It’s the same idea as the prisoner’s dilemma.
Imagine two prisoners are arrested for a crime and are interrogated separately, with each offered the following deal:
- If both confess, they each receive a 5-year sentence.
- If only one confesses and the other remains silent, the confessor goes free, and the silent prisoner receives a 10-year sentence.
- If neither confesses, they each receive a 2-year sentence.
The upshot is that, even though both prisoners would be better served if they both remained silent, each has an incentive to confess, since regardless of how the other prisoner acts, a confession benefits the individual.
(Full disclosure, I got this explanation by asking ChatGPT to “please explain the prisoner’s dilemma to me like I’m 5.”)
Discovery Mode works on similar principles.
If you don’t opt into Discovery Mode and everyone else does, you’ll probably get fewer streams. If you do opt in, you’ll probably get more streams – at the expense of the artists who don’t opt in.
But worst of all is that the more artists opt in, the less effective the promotion becomes for everyone, and the less Spotify pays out overall.
It’s an individual win at the expense of a collective loss.
With all of that said, I’ll be honest…
I’ve opted into Discovery Mode myself.
In my (ethically weak) defense, I did it months ago with virtually no thought for the moral implications of the program; I just make a habit of testing new promotional features, since I’m nearly guaranteed to be asked about them by clients.
Here’s how Spotify portrays my April results so far (which are pretty typical of my results each month):
I have a tough time taking all of this at face value. I mean, sure, these stats look impressive – a 328% boost in streams, wow! – but I don’t really see the numbers reflected in my total streams over time.
Take a look at my audience over the past 12 months, for reference:
The slow rise midway through last year is due to a) three new releases and b) Meta ads.
In February, I turned on Discovery Mode for five new songs. Despite the fact that Spotify tells me I’m now receiving an extra 10,000 streams per month from my commendable choice, it doesn’t seem like capitulating to the streaming overlords has had any real impact on my overall volume.
You can’t even see a bump when you separate out streams from Radio and Autoplay, contexts in which I should be seeing a Discovery Mode boost:
In other words, judging by my results alone, the program might be an overhyped load of crock.
For what it’s worth, I wouldn’t go quite that far.
I don’t have a wide enough collection of data to offer anything other than anecdotal evidence, but I have seen other artists get notable Radio bumps after opting in. So for some people, it seems to have worked.
But apart from the algorithm gods, I don’t think anyone really has any idea of just how effective the program is.
I’ll end things with three final thoughts.
First, I just wouldn’t worry much about Discovery Mode one way or the other.
Opt in or opt out, but don’t stress too much; it might get you found by a few more people, but it’s not going to make your career either way.
Second, don’t use the data for decision making. I’ve had a couple of artists ask me if they should put stock in how individual songs perform in Discovery Mode. They’ve seen lists like this:
…And they’ve considered using these metrics as a barometer to pick the songs on which they’ll focus their promotion.
I think this is the tail wagging the dog.
When I’m choosing a song to promote, I’d much rather focus on the track’s overall metrics (save rate, streams per listener, etc.) than cherry-pick weird stats from a narrower context. Shoot, I’d much rather just pick a song I really like.
I don’t think there’s much correlation between how much “Listener Lift” a song is getting and how well it would do in a targeted ad campaign.
Third (and last), I think if you expect Spotify to pay you well, you’ll nearly always be disappointed.
If you view the platform primarily as a means of getting your music heard, you’ll find it all less frustrating.
…A little less frustrating, at least.
To that end, I hope this email has been helpful, whether or not Discovery Mode ever is. As always, here’s wishing you good luck.
– Jon
