What Artists Should Know About Spotify Marquee
This changes everything.
Over the course of many posts, we’ve established that I can get potential fans from Facebook and Instagram to my latest release on Spotify for about $0.20.
Today I’m going to make a case for paying $0.50 to do the same thing from within the Spotify app instead.
Let me back up…
There are two ways to promote your music natively on Spotify: Ad Studio and Marquee.
I’ve written about Ad Studio twice, most recently in 2019. As a direct result of that post, Spotify flew me out to Brooklyn for an interview, parts of which made it into their artist education video series, The Game Plan.
I haven’t used Ad Studio since, but they recently emailed me an offer for a $125 credit. I’m tempted to try advertising my Vocal Synthwave Retrowave playlist or Synthwave Top 10 podcast.
Marquee is the new kid on the block, still in beta. To qualify, you need at least 15K streams in the US over the past 28 days, or more than 2.5K US followers. Additionally, your billing address has to be in the US.
The way it works is that when members of your “reachable audience” (more on that later) open the Spotify app, they’re greeted with a full-screen recommendation to check out your new release, like so:
If they click, you’re charged $0.50. If they don’t, you’re not. In both cases, Spotify tracks whether or not they go on to stream the release.
Album recommendations are presented to both free and premium users in the US. Single recommendations are only shown to premium subscribers, since a premium account is required to pick and choose individual tracks.
It looks like targeting will soon expand to more countries:
The link is to this form, which I just filled out myself!
Campaigns run for 10 days or until your budget is spent. My two completed campaigns, which I detail in full below, both ran 8 days.
Just like with Ad Studio, if the campaign isn’t living up to your expectations, you can pull the plug and only be billed for what you’ve already spent.
Spotify Marquee Campaign Creation
Campaign creation literally couldn’t be easier, in my case anyway!
At this stage in my career, the only thing I can tweak is the campaign start date.
The release is already selected and the start date defaults to release day, I think. You can create your campaign up to 18 days post-release.
Looking back on my campaigns, the two completed ones started the day after release, and my current one the day of. It may be better to wait until after your die-hard fans have a chance to listen, but I don’t have any data to support that hypothesis.
If more listeners were interested in my music, I could select specific segments of that audience: recently interested listeners, casual listeners, and lapsed listeners. For that higher degree of specificity, I’d pay $0.55 per click instead of $0.50.
As it stands, I only have the option of selecting my entire reachable audience, which was 2109 listeners at the time of my first campaign.
For my second campaign, it was 2856 listeners. My current one was around 4K (I forgot to take a screenshot). If I create a campaign today, I’m at 4418 listeners. Progress!
It strikes me that reachable audience is the key stat to track, more than monthly listeners (currently 25K) or even followers (currently 20K).
How Spotify determines the size and scope of that audience is a mystery. I assume it uses the same internal metrics that drive algorithmic playlists and Fans Also Like.
At 4418 listeners, I can still only select my entire reachable audience, which is fine. That’s what I’d opt for regardless.
Again, no real options here. I could enter a larger number, but it wouldn’t get spent.
Spotify Marquee Campaign Results
My first campaign was less than spectacular, for reasons I’ll explain below.
Since this was a single, it only reached premium users, because free users can’t pick and choose tracks.
195 people streamed the release after seeing the full-page recommendation, averaging 2.34 streams each for a total of 456 streams.
But that’s not all…
A quarter of those 195 listeners saved the track or added it to a playlist, demonstrating an intent to listen to it in the future.
Spotify also reports how those listeners engage with the rest of your catalog, which is arguably the whole point. We want fans, not just pricy streams.
In my case, 34 listeners generated another 200 streams. That’s 650 streams total, which will eventually pay me $2.50 at $0.004 per stream, bringing the true cost of the campaign down from $118.50 to $116.
It stands to reason that these campaigns are unlikely to pay for themselves.
To be fair, this particular release wasn’t an ideal test subject. It was a double A-side containing two covers that were already released as part of my Depeche Mode tribute reissue a year and a half prior. I didn’t announce it to my fans or otherwise promote it.
Worse, a key synth part somehow didn’t make it to the final mix of “Sometimes.” I can’t overstate how devastating this omission was, right at the climax of the song (you can hear it at 2:09 here).
I’ve since replaced the track on Spotify, which sadly required a new ISRC. I lost my stream count and playlist placements in the shuffle.
Let’s move on before this gets seriously depressing!
My second campaign was for a follow-up EP to my last album. It contains 6 b-sides and 5 remixes for a total of 11 tracks.
Since the release is categorized as an album, the campaign reached both free and premium users. Free users can only listen on shuffle.
Perhaps that explains the 4x improvement in streams per listener. Free users may be holding out for specific tracks, like the opener.
As you survey these substantially larger numbers, keep in mind that the campaign cost twice as much!
That’s 371 listeners x 9.73 streams per listener = 3610 streams.
The conversion rate went up by 2%. Maybe the artwork was ever-so-slightly more appealing, or more listeners made it past the 30-second mark, or both.
A modestly higher intent rate generated “now you’re talking” levels of saves and playlist adds.
This campaign generated fewer playlist adds and saves on my other releases, but more listeners and streams. 49 listeners x 11.92 streams per listener = 584 streams, to be exact.
That’s 3610 + 584 streams = 4194 streams total x $0.004 per stream = $16.78 I’ll eventually get back. Whoopee.
Spotify Marquee Conclusions
If I had to choose between Facebook Ads to Spotify or Marquee (and I kind of have to since I’m not made of money), I’d choose Marquee.
With Facebook Ads, I average 50 conversions (clicks to Spotify) at $10 per day. Assuming one stream per click, that’s $280 for 1400 streams over the course of 28 days.
Compare those 1400 streams to this:
Spotify gives me over 8K streams for free, precision-targeted to the users most likely to enjoy my music.
And that’s not even factoring in geography! Most of my Facebook Ads conversions come from Mexico, Brazil, and previously, Russia.
In contrast, two-thirds of my algorithmic streams come from the US:
Spotify introduces new listeners to my music who I retarget with Marquee to grow my reachable audience.
That’s how I think it works anyway. I have no other explanation for how my reachable audience went from 2109 to 4418 listeners in just over two months, with no substantial change in monthly listeners or playlist placements.
Then again, the two campaigns only converted 566 listeners while my reachable audience grew by 4x that number. And that’s not even factoring in any overlap between the two campaigns.
Marquee has additional advantages over Facebook Ads:
Implied endorsement. When Spotify recommends your new release, it feels more like a suggestion than an ad. In reality, it’s both! Marquee only reaches users likely to be receptive to your music.
Deep, useful data. We finally have a way to reliably measure fan growth on the platform and the relative performance of each release. “Reachable audience” may just be the one metric to rule them all.
Marquee also has disadvantages:
US-only. If you’re not in the US, you’re out of luck, for the time being anyway. I look forward to the day when I can target Brazil and Mexico at a lower cost per click.
Walled garden. There’s no reliable way to engage with your reachable audience beyond additional Marquee campaigns. These “fans” aren’t attending shows, buying merch, or otherwise supporting you in any financially meaningful way.
For now, I’m going to continue promoting my playlist with Facebook Ads and my new releases with Marquee.
If that gets too expensive, I may need to restrict my Marquee campaigns to album releases.
Have you tried Spotify Marquee? Share your thoughts and questions in the comments!
I like your recommendation! Thanks and I wish you continued success with your releases!
Much appreciated Dave!
Finally the post I’ve been waiting for! For artists with basically no US following, would you suggest running FB ads to only the US for a month to get the required “15K streams in the US over the past 28 days, or more than 2.5K US followers” and then run Marquee ads? My US fan base is nearly non-existent and I’m scared to run expensive US ads for a whole month just to be able to use Marque when I KNOW $250 will get me 400 listeners/4000 streams/120 Followers from “lower tier” countries using FB ads.
I realized after posting that the entry requirements would be out of reach for most of my readers. Without Spotify’s algorithmic help, I’m not sure I’d hit 15K US streams in a month!
The danger of running Facebook ads to hit those numbers is in not quite reaching them on a month-to-month basis. Granted, once you hit 2.5K US followers you’re probably safe.
My only other thought would be to get on playlists with predominantly US listeners, but I’m not sure how you’d find them unless they’re like… patriotic playlists? Like “4th of July Bangers”!
This Marquee strategy seems better than facebook (which we’ve never used) but… it still seems pretty crap based on the numbers you provided. It only looks good in relation the other strategies, which are downright horrible. On Spotify, you really want to reach (1) subscribers, who (2) like your genre. And there is really no remotely efficient way of doing so through advertising. What would be helpful is if Spotify took some unobtrusive, small corner of the account screen and labeled it “Check out,” which would have links to new releases, or whatever, by artists who fit the user’s genre and paid for these placements. It’s not perfect but… that would be FAR better than anything Spotify’s come up with so far. Right now, you’re largely limited to non-subscribers who, by their very nature, aren’t going to be very valuable.
Spotify’s algorithm does essentially what you’re asking for free by inserting our tracks into their playlists. Maybe I’m misunderstanding your description, but it sounds like a pay-to-play Release Radar.
As for Marquee, you already reach (1) subscribers (premium always, plus free with album campaigns), who (2) like not just your genre, but your actual music. That strikes me as ideal targeting.
I’d wager that fans of your genre who haven’t heard of you are far less likely to click through than people who actually know your music.
Yeah, I think a pay-to-play release radar is a good way of putting it. It would kind of give you a wider reach than just relying on the algorithms alone to get onto the release radars you’d like to be on. (And I should have said “premium subscribers” above to be clearer.)
Again, Marquee looks better than most other options but… still not worth it in my eyes based on the numbers you posted. But that’s somewhat subjective, of course. And others’ results might be different. And everyone’s tolerance for losing money on a particular measure regarding their hobby is different.
Arguably Spotify has eliminated hobbyists with their relatively steep requirements, but your point is taken. For now, I feel like Marquee is the best option for Spotify promotion, but that could change if/when it gets too saturated and users get sick of seeing a pop-up every time they open the app.
I think you mean that Spotify has eliminated hobbyists from using Marquee. But I’d argue that unless you’re making a living, or at least some meaningful net income, from songwriting… you’re a hobbyist. And just 16,500 of over 8 million creators on Spotify generate at least $50,000 in streaming revenue annually – that’s 1/5 of 1% of all creators. After expenses and taxes that number’s probably down to $20,000 or less. My point being that 99.5%+ of Spotify creators are hobbyists, including you, me and just about everyone who comes to this website.
By your definition, Spotify itself was a hobby up until this year, when it finally became (slightly) profitable. Pretty sure we’ve talked about this before, and in the past I may have considered music a hobby, mostly to take the pressure off.
Obviously I’d bring the goalposts considerably closer. I think it’s fair to consider anyone meeting Spotify’s Marquee criteria to be a professional musician.
Of course we can push too far in the other direction as well, if we simply define “professional” literally, as simply earning (any) money from it.
Spotify is a business that employs many thousands of people, all of whom are earning a profit on said employment. And owned by shareholders who have earned a massive profit (so far). And that was before they ever earned a profit. So, definitely not a hobby.
You’re a professional when it comes to mixing and mastering music. That’s your job and you get paid to do it. But you’re a hobbyist when it comes to songwriting; it’s an avocation, not a vocation. Like *almost* everyone else who writes songs. But I realize some folks have a certain self-image of themselves that’s important to them, and that’s ok.
I wish they’d show how many of the people who saw the ad followed. What would you guestimate?
I have no way of knowing. I get a guaranteed 300 followers a month from Rise so I can’t tell one source from another.
Considering I have way more followers than my “reachable audience,” perhaps the former is already fully contained in the latter.
Hang on … Rise? Have a missed a write-up on them?
Nope! I’ve been promising one for a long time, but they’ve assured me a revamp is on the way, and I don’t want my screenshots to be immediately outdated. I’ve talked about it several times in my email newsletter though!
“I assume it uses the same internal metrics that drive algorithmic playlists and Fans Also Like.”
This makes me a bit scared. “Fans also like” hasn’t been updated in years. With 1k playlist features and new followers you’d think “fans also like” would update. It seems Spotify keeps things very mysterious. Knowing your audience and followers would be very nice, but it’s kept a secret.
My Fans Also Like is eerily accurate, but I’m not sure when it last updated. I’ll keep an eye out for when my latest collab partner shows up! My previous one isn’t there, which is suspicious.
What do you mean collaboration partner? Spotify blogs about looking at what the artist is doing outside of Spotify as well – it doesn’t make sense. Fans also like not been updated in 8 years. I’ve received mail that I wasn’t supposed to receive, sounds weird right? It shows how they work.
I’m referring to the other artists I’ve collaborated with. My current collaboration partner Minute Taker isn’t in my Fans Also Like yet, but I see that as of today, I’m in his! Our collaboration came out on Friday.
Perhaps Spotify just isn’t getting a clear enough signal as to who your true fans are?
FAL is a weird thing… ours was crazy inappropriate for a while when we first got on Spotify, then after about 6 months it became, as you described, “eerily accurate” – I mean, spot on. For about a year. Then, poof, it just disappeared for no identifiable reason. We had a recent release that’s gotten on maybe 50-ish playlists (two Spotify algorithmic), all but three of them right on the money genre-wise and still no FAL. (We have ~15,000 monthly listeners.) We’ve also had a ton of reviews which, according to Spotify, some of which should be picked up by the algorithm to influence FAL but… zippo. Personally, I’d rather have no FAL than have it incorrect – it’s not a big deal – but… it would be better to have a correct one than not have it at all.
A huge proportion of Spotify creators have no FAL – I’d guess 80% or more – because they have so few listeners that the algorithm can’t do its job. 93% of Spotify creators have fewer than 1,000 monthly listeners. So you can back into that math. But no or inappropriate FAL is a very common issue as (also) evidenced by the number of creators that have questions about it (you can google that). If the vast majority of creators either don’t have a FAL or it’s screwed up… that seems like something Spotify would want to remedy.
My suggestion – which will never be implemented – would be to change the name to “You Might Also Like” and the default setting would be 100% Spotify algorithm, but… creators would have the option of choosing a setting in which they chose a certain number of the artists (8?) for this section on their own. That way, worst case, the creator has a section, albeit small, that they picked themselves. I can’t think of any downside to this whatsoever. And it can’t be a very complicated fix from Spotify’s perspective. But, again… it’ll never happen.
No FAL is certainly the preferable option! Are your Radio selections also skewed? Seems like it would rely on the same logic.
I hear where you’re coming from with “You Might Also Like” but I guess “Fans Also Like” makes it clear that the suggestions are based on the selected artist, whereas the former might be interpreted as general recommendations based on historical behavior.
Bizarrely coincidental… our Fans Also Like section reappeared the day after I posted this. And while not eerily spot-on like it was in the past, it’s genre-specific and all of the components make good sense. So I guess the algorithm is working on our account again. Strange timing, though.
Also odd is the fact that I wrote to Spotify after our first FAL showed up a couple of years ago and it was totally screwed up. I got a reply explaining that it was out of their hands and algorithm-driven (which I didn’t know at the time). Ok, fine. Then, within a week, it changed completely and was dead-on. Coincidence? I guess.
It’s all a bit suspicious but… whatever. It appears to be working now.
We know Spotify searches the web for mentions of bands to help place them properly. Maybe they also search for FAL complaints. 😉
Dope article. So are you saying you’d value spotify marquee at 0.50/stream even tho that’d be more expensive than bringing a fb/ig user to your spotify, because spotify marquee is specifically targeting users who already like your music (reachable audience) or because using marquee gets your more expansive date ie “Spotify also reports how those listeners engage with the rest of your catalog”
I didn’t fully grasp how this broader data is provided or what it might open up, so I thought I’d ask
Thanks as always
I can think of lots of reasons to value Marquee over Facebook Ads. Most importantly, we know the stream actually happened! With ads, all we know is that they clicked to go to Spotify.
Beyond that, there’s improved targeting (to the people who like your music, not just people who are likely to convert on Facebook) and deep analytics you mentioned.
As for how it’s provided, what you see here in the post is what you get. I’ve screenshotted all the data.
Thank you for sharing this review about Marquee. I’ve seen Andrew mention it as well as a few other buddies, but haven’t considered it an option quite yet.
It does bring in a little more traffic for the amount you are putting into it.
Have you considered running traffic over conversions? At $10 a day with traffic and only tier 1 I was averaging between 29-40 conversions at a cpc of .13-.18. Still not great I know, but from experience, it doesn’t look like conversions pull in the weight they did before IOS14.
Thanks again for the post. Very insightful!
I was actually considering running traffic for my playlist. I’ve done it before in the past on my artist profile and it ran off the rails more than once, to the point where I had to ditch my lookalikes because they were corrupted by bots.
I also noticed that unless I omit YouTube from my Smart Link, a lot of the traffic ends up there. Which is fine I guess, but I prefer to optimize for Spotify.
Great intel, Brian. Thanks, as always!
And sorry about “Sometimes.” Sometimes, these things happen…
I’m not laughing…you’re laughing… 😉 (I kid, of course. Sorry. Serious bummer. I had a similar thing happen once. Would’ve had to start from scratch to re-do, etc.)
I definitely spent some time kicking myself over it! To be fair, I can’t expect myself to listen all the way through every render out of my DAW. For some reason one of the softsynths didn’t load the patch, or was the wrong version, or wasn’t M1 compatible, or some such nonsense.