The 5 MOST Useful Spotify For Artists Metrics
Unlock the true power of Spotify for Artists by exploring five essential metrics that reveal how your marketing campaigns impact streams, saves, and playlist performance.
Playing a video or selecting a transcript timestamp connects to YouTube, which may use cookies. See our privacy policy.
Music Marketing Monday
Get the written breakdown of ideas like this every Monday.
One practical idea a week, read by 18,000+ artists, managers and labels.
Quick summary
This video dives deep into advanced Spotify for Artists metrics beyond basic stream and listener counts, showing how to analyze active versus algorithmic streams and the importance of playlist placements. Understanding these details helps artists measure the effectiveness of their marketing efforts and identify which sources drive meaningful engagement and saves. By segmenting streams by source and examining listener behavior, you gain insights into long-term audience building and campaign impact. The video also highlights comparing song performance over time to gauge growth speed and engagement quality, offering artists a premium approach to optimizing their Spotify strategy.
Transcript(English)
Hey everyone, this video I'm going to go over how I typically analyze performance inside of Spotify for Artists to get a good feel for how a given marketing campaign I'm running is going because I've noticed a lot of people don't seem to know how to navigate Spotify for Artists and extract useful information outside of just like raw streams, listeners, save metrics. And also talk about how this can work for different scenarios, different types of campaigns you might be running, tracks versus albums, etc. So you can dive a little deeper than just my song has 2,000 streams, it feels like it's good. >> [music] >> So diving into Spotify for Artists, we're going to start off actually going and looking at a given song just because I think most of you are probably looking for song stats and analysis first. Uh this song's been out for a little bit
more than a month. If I look at the graph, you can see this here. Um so we're just going to kind of ignore the couple days before the 28-day window just to make this easy. Now, of course, when you come to Spotify for Artists, you get all these basic stats. And if you want to see kind of like a just base level how to jump around Spotify for Artists, I do have a full-length video here you can see. Um but chances are if you're watching this, you already get how to get the basic info. So you have streams, you have listeners. Listeners are the unique people that stream the song. And so you have a person listen 1.75 times on average. Um and that's how you end up with more streams than listeners. People can add songs to playlists and people can save songs to their library. And um that's all great and that's super useful information when you're trying to gauge your campaign. But you can go a little bit deeper. For example, one
thing that I do super often, like it's one of the first things I do when I'm looking at a song nowadays, is I go to segmentation source streams and I look at active and compare it to personalized playlist autoplay mixes, aka algorithmic. So I put at all active and then personalized playlist autoplay mixes. And this gives you an immediate feel for how the song is performing algorithmically. Um because the the blue here is your actual active. So, if you're running an ad campaign uh or you're doing a social media campaign and probably both, hopefully. The blue is your active and the pink here is your algorithmic. And if you're wondering like the types of ad campaigns I'm talking about, I'll link to a video right here where you can see how to run like a meta conversion campaign from start to finish. But but that's what this blue is here. And so by looking at
the blue versus the pink, you get a feel for like is my is the algorithm pushing out my music and when is it pushing my music, how much is it pushing it out. In this case, the song has actually gotten more algorithmic streams than active streams. So, this is like a very good result, right? Especially this big spike here. If I take off the source of streams, like you see the spike and you're like, "Oh, wow, like what's happening?" You don't have the full context of to what's really happening on this day without looking at segmentation source of streams. But by looking at active and then algorithmic, it's obvious. There was an algorithmic spike on Monday, September 7th. And you know, knowing what I know about the algorithm, I know that Monday is generally when Discover Weekly updates its playlist. And so I can come up to playlist here, which is another I guess kind of trick to analyze songs. And I do see Discover Weekly. Now, I can
Show full transcript
go a little deeper cuz I don't know if that spike was definitively Discover Weekly. Because it's the current day, I can actually just say last 24 hours. Then I see it's 1,153 streams from Discover Weekly in the last 24 hours. And if I compare that to last 28 days, that's almost all the Discover Weekly. So, what this means is that spike is basically we got a really good Discover Weekly push on that song, which which is very exciting. Um But also coming here, it it's gives you a good feel for how you perform in different algorithmic playlists. For the most part, most people do best in radio. Recently, I've seen a couple people where mixes was their best one. And for a lot of people, I'd say your DJ and smart shuffle are fairly low. But I've seen some people where your DJ just does amazing. So, I think there's some genre
differences here, maybe some demographic, cultural differences, etc. Um but radio is typically number one, mix is number two, and then discover weekly and release radar number three and four. That's typically kind of the order here. Radio is the the big kahuna nowadays. So, keep that in mind. Now, going back to source of streams here, if you have playlisting going, you can instead look at active algorithmic and then other. And then this will also this green line here is is the other listeners playlist or third-party playlisting. In this case, the song doesn't really have any. Um aside [snorts] from my own playlist that I own, which shows up as this green. So, but if you're running a playlist and campaigning, this is how you can kind of say this is how the playlisting's going separate from the ad, separate from the organic, which lets you gives a pretty clear picture onto how maybe different companies you're hiring performing or or
whatever it is, you know, just trying to get an overall feel for how you're doing. Going back into source of streams, um you can also split active into artist profile and listeners on playlist and library. And generally what I'm looking for here is to get an understanding of how I'm building long-term listeners. So, the the purple whatever color this is, pinkish purple, is essentially like first-time listeners. So, this is still my ads and still my social media numbers coming here. Um but the the green is people from those same places, but their second listens, right? So, if someone comes directly to my song from my ad or social media, their first play is going to be from this pink line. After they
save the song or add it to a playlist, those future streams from those places are going to show up in the green. And what happens is the green kind of slowly grows over time, right? You can see this here. Um and the even though the pink's staying pretty flat. This is because we're kind of building like a a almost like a list of people who are going to listen to our song on repeat because they they've baked it into their playlists, they've baked it into their libraries through saves, and um that's why it's kind of good to get an understanding for how the first-time listeners are converting into people that saved and playlist added to a playlist and then are re-listening because when you turn off your ads that this pink will will pretty much immediately just crash, but this green
is going to be what gives you that tail uh in addition to algorithmic. So, if you kind of combine listeners on playlist and library and then personalized playlist, this is basically like what's going to give your song life after you turned off the the actual ad campaign or whatever marketing campaign you were doing. The ads, they're not going to send new people to your song after the ads are off. The playlists are eventually going to take your song off. So, this is really like what's going to keep your song moving. Now, another kind of hack that I often encounter is people wonder like how is my song performing in terms of save rate, but also how is it performing in terms of like streams per listener actually? Because what can happen is if you have a song that a really good algorithmic push, algorithmic playlists have a very low save rate and a very low
stream per listener rate. So, what happens is the the stats engagement stats look very bad, but that's just because playlisting of any kind, whether it's algorithmic, editorial, or third-party user playlist, it all does very bad in terms of those metrics. So, how good are your actual ad engagement engagements going? And the way you do that again is source of streams. Um active, algorithmic, other, right? And then I can look here at streams per listener, for example, and then down here we see um active sources are almost four streams per listener, which is quite good, right? And then we see our typical kind of playlist stats, 1.27, 1.6 1.6 for other listeners playlist is actually quite good. 1.2 1.3 for algorithmic is actually also quite good. I've seen 1.1 here pretty often, actually. So, these
are healthy stats, but it lets me know that like, "Oh, hey, the people I'm sending are actually listening quite well. So, it's not that my stream per listener is poor here, right? This looks very bad. It's because the song's doing so good algorithmically that the stream per listener is bad. Now, on the similar note, I can go to listeners here. If I come back down, um we can see how many listeners we're actually getting from active sources. So, we've had 2,120 listeners from active sources, and we've had 1,992 saves. Well, I know from experience that other listeners playlist gives you very little saves. Algorithmic gives you very little saves. Even with this massive spike from uh algorithmic, if we look at saves, um the the spike in saves was pretty marginal, despite there being an extra 1,100 streams on that day, right? Um so, that should
hopefully give you proof in long, but I I've seen it across dozens, hundreds of other songs. Um so, if we come and split up these three sources again, right? We know that in terms of listeners, which is what I'm looking at now, we had 2,100 listeners, and we had 2,000 saves. So, that tells me like pretty much all these saves are from the active people. So, my save rate is like 90 5% something like that, right? Almost 2,000 out of 2,100. So, it's it's quite literally 95%, which is awesome. Um and also, most of the playlist ads are going to be from people who are active listeners. The playlist ads do You do get some playlist ads from different playlists as well. Um but you know, a good chunk of these are that. So, what this tells us is it's not that
like our save rate is poor because only um 2,000 out of 13,000 listeners saved it. It's that algorithmic stuff and whatever doesn't give you many saves. So, this kind of helps you understand how good is your ad campaign actually doing in terms of engagement. Now we know, right? So, now we know specifically we're doing really good algorithmically for this song. We have more algorithmic streams than active, which is phenomenal. And we actually have a very, very good save rate, a very good stream per listener rate uh for the song only being a month old, and a very good playlist add rate. Whereas, if you were just looking at these core stats alone, you might actually even be disappointed with how people are interacting with your song. Um but you would have a false understanding of it. So, another trick I want to show you here that's related to this is that when it comes to comparing songs, a lot of times
people are wondering like how fast did it take me to get to this many streams? Um you know, I didn't do marketing for the previous song, but I'm currently doing marketing for this song. How are things comparing? You could of course flip back and forth, open two browser windows, but there is a pretty handy tool here called compare songs, where you can look at like the first 7 days for the the some previous song. Our previous release was an album, so um this isn't an ideal comparison, but for uh for for example, this song "Cold" when it came out in the album, and again, you want to do this with two singles, not with an album, but it it it overlays the first 7 days for for both, right? So, I can see that the blue, which is my car Oh, I got to scroll down. I can see that the blue um compared to my the dotted line here,
overall, I'm beating how good this song was when it came out. Um and if I go and change this song to a different song, maybe I do first 28 days, and in this case, I look at this song that actually did come out when the album came out, but this song we didn't like promote dedicated. Of course, we're crushing it, right? But this works very good when you're comparing two singles, because it's just kind of an easy way to tell like, "Oh, yeah, like we're drastically shooting shooting past what we were getting early on." Um and you can of course compare stream per listener. I don't remember if you compare saves. You can. You can. Which is good. So, so you can compare all these metrics here. You can't segment them, though. So, if you if you want to do like segment comparison, like how good was our algorithm in the first week, you'd have to put them side to side. But this just kind of an easy, quick way to get a feel for like
like oh yeah, we actually are doing like 25% better than our previous one and we got a bigger algorithmic spike one week into the song where the other song was later and the save rates actually way better, right? So all these little things um you get by doing this kind of quick comparison that you don't get otherwise. And this next one's a really deep one. So for example, let's say you're running a campaign, you can't really tell what it's doing, but you looked in your ad campaign and you saw that like oh I'm getting a lot of conversions in Brazil or the US or the UK, whatever it is, right? And so you want to see, you know, I'm getting all these streams, but I don't know what's coming from my ads, what's coming from elsewhere, like maybe you already have 50,000 monthly listeners, so you already get pretty good traction. So you don't really know what your like $20 a day ad campaign is doing if if it's not doing like really well, it's not making a noticeable
change. So what you can do is I can go to filters and let's say I know that it's it's coming from the United States, I can come here and type in United States. Now all the data here is filtered for the US, so I can look at saves in just the US or I can go to streams source of streams active versus algorithmic and now I'm looking at a segmented source of streams for just one country which is like, you know, you're stacking all these different levels of analysis, but the purpose of this is um you know, for for example, like I can use this to tell um that on Monday when I saw this algorithmic spike, we got 372 algorithmic streams from Discover Weekly, right? We know this spike was from Discover Weekly and it was about 1,100 and about almost 400 of them are from the US, right? So we couldn't have
known this without using this stack of uh countries and segmented source of streams and it also lets us know kind of like how much traction we're getting from socials and ads from there and then you can of course flip the the countries around to look at individual countries to try to get overall understanding of like oh we spent a bunch of money in Brazil and Mexico, we get a bunch of streams and the algorithm is in the US, which is actually a super common scenario, but you can prove it here, right? I I We've all known this for years, anyone that's been running campaigns like this for years, that it doesn't matter where your streams come from most of the time, the algorithmic spike you get is very often in the US or specific other tier one markets. Like I've had some in the UK and some in Germany. Generally for me, most of my algorithmic pushes have been in the states. So, this is how you actually prove that. Now, if you want to see how
you can run campaigns like the ones I talked about in this video, you can check out this video right here to see that entire process from start to finish. I hope you enjoyed this nerdy look inside of Spotify for artists and how you can use it tactfully, tactically, either one. And if you did, please make sure you subscribe and ring that bell so you don't miss my future uploads. Thanks for watching. I'll see you in the next video. Bye.
Grow your music with a proven ad system
Learn the Meta ads and streaming growth playbooks Andrew uses across artists, labels, and agency campaigns.
Keep exploring