Facebook Ads for Spotify Best Practices

Disclosure: This post may contain affiliate links, meaning I get a commission if you decide to make a purchase through my links, at no cost to you.

I’ve written about Facebook and Instagram ads for Spotify many times, like here, here, and this one from April of 2018 where I noted that I’d spent $5888.35 on Facebook ads to date.

Three years later, it’s $20,614.84.

While there are still useful angles to explore in those older articles, this post supersedes them. Not because they were necessarily wrong, but because the platform and practices have evolved.

For example, Instagram Stories was far and away the best ad placement for music two years ago. Viewers tend to be receptive to exploring new music, and have their sound on.

The problem is, most musicians know that by now. The resulting competition has driven ad costs up. Lately I’ve gotten cheaper followers through feed ads.

My Spotify following in March 2021

I’m spending $20 a day on Facebook ads, split between promoting my artist profile and my synthwave playlist.

I’m up to 9K profile followers and 17K playlist followers.

The big bump in the middle came from my Gleam ROLI giveaway

Every morning I track my follower growth in a spreadsheet. Here’s March so far (click to enlarge):

conv = conversions, CPC = cost per conversion, CPF = cost per follower

Right now I’m paying about $0.30 per follower for both my profile and playlist.

The latter has been pretty stable, but the former is just starting to settle in since I added India last week. More on that later…

My current Facebook ads for Spotify

Here’s my current profile ad for feeds:

And my playlist ad for feeds:

Same audio! For whatever reason, this little snippet of the last verse going into the coda outperforms everything else I’ve tried.

You may have noticed my feed videos aren’t square. They’re actually 4:5 or 1080×1350. I render them at 2x (2160×2700) in ScreenFlow for retina screens, though I’m not sure it makes a difference after Facebook processes it.

The result takes up more screen real estate on mobile, at the cost of black bars on desktop:

I get very few conversions on desktop, even with square video, so it’s totally worth it.

My stories ads are 9:16 (2160×3840), and are basically the same as the feed ads but with an added prompt to “swipe up to follow.”

I run both variations in a single ad using the “edit placement” function:

The eight placements I’m running are Facebook News Feed + Video Feeds + In-Stream Videos, Instagram Feed + Explore, and all three Stories options.

It doesn’t really matter though. As long as you don’t waste your money on Audience Network, Facebook will optimize for the best performers.

My profile ads dominate in IG feeds, but my playlist ads do better on Facebook (click to enlarge):

Breakdown by Placement, March 1-20

Facebook Ads for Spotify Targeting

I target all Spotify countries, divvying them up between my profile and playlist ads so they don’t compete against each other in the auction.

The playlist ad hits the Americas plus Oceania, which roughly translates to Mexico and Brazil (click to enlarge):

Spotify playlist ad, Breakdown by Country

The profile ad delivers to the rest of the world. Mostly Russia, but India will surely overtake it soon:

Spotify profile ad, Breakdown by Country

Up until last week, my ads wouldn’t run to India, and I couldn’t figure out why until Facebook support told me:

Facebook page restrictions

At some point, I must have blocked a bunch of countries from viewing my page’s content.

Since removing all country restrictions, most of my profile conversions are coming from India. I’m good with that! Spotify for Artists confirms that they’re actually listening to the track:

Can you spot when my ads to India started delivering?

That’s not always the case though. The $6.55 I’ve spent so far this month on Algeria doesn’t appear to be paying off, so I may drop it:

PSA: Country comparisons are only available for individual songs!

I’ve already dumped Ukraine, Kazakhstan, and Indonesia for the same reason.

Considering where my ads are being shown, you’d think that my Spotify Top Countries would be Mexico, Brazil, and Russia. Nope! I have more listeners in the US than those three countries combined:

USA! USA!

By now you’ve probably heard of the whole “trigger cities” concept. By advertising in cities that have high music consumption and low advertising rates, you trigger the Spotify recommendation engine, resulting in more algorithmic streams worldwide.

I’ve never zeroed in on specific cities, and I can’t prove that the underlying assumption is correct, but I can testify that 15% of my streams, 8.5K in the past 28 days, are from algorithmic playlists.

Conversions Objective Setup and iOS 14

I’ve described where I’m targeting, but not who. I’m targeting a lookalike audience of users who previously clicked through to Spotify from my ToneDen landing page, which looks like this:

The Spotify link directs to my profile, not the song

I optimize for Spotify clicks using a custom conversion:

Facebook Ads custom conversion

You might be wondering, don’t I want YouTube and SoundCloud plays too? Sure, but I don’t want Facebook to optimize for them! I can already get YouTube views for far less.

Explaining how to set everything up in the midst of iOS 14 changes would take its own quite boring post. Before you can configure your ad, you have to verify your domain, create the custom conversion, add it to Business Settings, and configure the web event in Aggregated Events Management.

And before all of that, you have to host the link on a domain you own, which took me five minutes with ToneDen. Just go to Settings/Advertising, click on “Add Domain,” enter a subdomain, then add a CNAME DNS record at your hosting service:

Now all my go.colortheory.com links direct to ToneDen. You can do the same with Feature.fm and Hypeddit.

If this all sounds too technical, I recommend John Gold’s Spotify Growth Engine. It’s dirt cheap through my link, and John hosts a weekly coaching call and Facebook Group where you can ask questions if you get stuck.

So again, I’m targeting a lookalike audience of users who already performed my desired action: clicking through to Spotify. Of course, I didn’t have that audience to begin with, so I started by targeting a lookalike of 75% video viewers.

My only other restriction, which is hardly a restriction at all, is setting the age range to 16-64.

So that’s my current playbook, honed from months of careful testing. In the near future, I’ll swap my profile and playlist targeting and retarget users who watched 50% of the prior ad. In other words, I’ll show the playlist ad to users who already watched half of the profile ad, and vice versa.

Beyond that, I’ve got a laundry list of further ideas to try out, and a list of conclusions I’ve drawn so far, which I’ll expand on now.

Facebook Ads for Spotify Dos and Don’ts

Without further preamble, here are my best practices for Facebook and Instagram ads to grow your Spotify following:

DO visit Facebook Ad Library for inspiration

Did you know you can view anyone’s ads? You should absolutely explore how other artists in your genre promote their music.

If you see a blank page, click on “All Ads.” It defaults to politics for some reason, for me anyway.

DO keep your ads simple and direct

I’ve tried animations, Spotify canvases, and official music videos. None of them work as well as the simple images you see in the examples above, with subtle animated overlays courtesy of the Motionleap app on my phone.

DO include the Spotify logo

The Spotify logo instantly tells users what your ad is about, and lessens the chance that they click through just because they like the pretty picture. You only want to pay for, and optimize for, Spotify users.

You can grab the Spotify logo as a transparent PNG from me here.

DO use Dynamic Creative to test headline, text, and description

Believe me, you can do better than “Out now on all major platforms!” I’ve found that “Listen now on Spotify 👉🏻” is a solid choice for a description. If Facebook doesn’t think displaying the description will help, it won’t show it to the user.

Oftentimes the best text is no text, especially in countries where English isn’t the primary language. You can test it by leaving the first option blank.

When in doubt, tell users what you want them to do, i.e. “Listen & follow on Spotify.”

DO experiment with Detailed Targeting Expansion and Lookalike Expansion

Lookalike Expansion always seems to help, even when it pushes the audience size into the tens of millions. Still, I don’t enable it right away, because I don’t want to risk optimizing for bots.

Detailed Targeting Expansion comes in handy for broadening a very narrow audience. Here’s one of my favorites:

For this campaign, I’d also throw in “And Must Also Match: Spotify.”

Depending on your age, gender, and geographic targeting, your potential reach could be in the low thousands or worse. Detailed Targeting Expansion can bring that up to a reasonable range, ideally 1-3M.

Again, it’s best to wait until you need it, like when the ad frequency gets above 1.5 or so. If your audience size is already workable, there’s no need to risk showing ads to users who may not have Spotify.

DO test different video lengths

Stories ads are limited to 15 seconds, but your feed ads can be as long as you want. I didn’t see a big difference going from 20 seconds all the way up to a minute, but I suspect it depends on the song.

I suggest ending the ad on an unresolved chord or phrase. It may encourage users to click through to “close the loop.”

DON’T bid against yourself in the auction

If you have multiple ad groups with overlapping audiences, you’ll bid against yourself in the auction, and your cost will go up.

The way to check is to hover under the ad set name and select “Inspect”:

If you don’t see 0% auction overlap, you’re doing it wrong.

An easy way to remedy overlap is to exclude every ad set’s audience from the rest.

For example, if you’ve got one ad group targeting Depeche Mode fans and another ad group targeting Electronica, you’ll almost certainly have overlap. You can see exactly how much in Audiences by building the two audiences and selecting “Show Audience Overlap.”

To remove the overlap, you exclude Depeche Mode from the Electronica ad group, and exclude Electronica from the Depeche Mode ad group.

DON’T forget that social proof is a factor

My ads have hundreds of likes, dozens of shares, and a handful of nice comments. That social proof is enough to outperform a 10% better ad.

If you’re always testing new ads, you’ll never reap the rewards of social proof. Sometimes the best ad is just “good enough.”

This works both ways. One insightfully mocking comment can torpedo your entire campaign, so be sure to keep an eye on your feed ads on both platforms, and reply to comments.

DON’T violate Facebook policies, even when you know you’re not

For years, my profile ad used the same image as my playlist ad. One day Facebook decided it was “shocking.” I appealed and they upheld their decision to not let me run the ad.

I could’ve made an imperceptible tweak and uploaded the video again, but I decided it wasn’t worth the risk.

A friend’s entire ad account was recently disabled for no reason we can fathom, using the same techniques and structure I outline here. No ad was ever rejected, no policy ever violated.

Still, the appeal was denied, and a half dozen requests for clarification yielded nothing but canned responses.

My advice is: don’t push your luck. Even if you know you’re in the right, assume that you won’t be able to make your case to a fellow human being.

Now let’s get out there and waste some money!

Is it smart to throw our money into the data black hole that is Spotify?

Probably not. Are we going to do it anyway?

You bet we are.

I’m spending $600 a month on ads to Spotify, and getting $300 back in royalties. I’ve built a finely tuned money-losing machine!

To be fair, I’m also getting 150 Facebook and 150 Instagram followers every month.

Sure, I could be devoting that $600 to selling CDs or running another free + s/h funnel, but growing my Spotify audience is addictive.

Let’s start with the obvious. Tens of thousands of people are hearing my music every month on Spotify. That’s too many for my brain to keep track of. It’s literally mind-blowing!

It’s also incredibly gratifying. That’s the whole point of making music, right? For people to hear it.

To the listening public, an artist’s monthly listener count is the primary measure of their success. If it’s too low, many won’t listen in the first place.

I don’t know who these Spotify listeners are and I have no way to reach them, yet I can watch my listener count on my phone in real-time.

I don’t know if Spotify gamified their artist platform intentionally, but it sure feels like it!

And so, against my better judgment, I persist in the Spotify ads game. I’ll continue to report my results so that you can do the same.

What’s working for you? Anything you’d like me to test? Let’s strategize in the comments!

43 Comments

  1. I found that my spotify marketing had an ROI about 16 months out.

    Something I did that not everyone can do, is I built a web-app which is the landing page that they click though to Spotify to playlist the song, yes they have to authorize connecting and yes I saw about a 60% drop in click throughs to Spot. But what I got back, using their API, was all the pertinent data about them – email, name, dob, location and sub level. The landing page had a bit more of the song, and basically said click to add to your existing playlist or create a new one. (and by clicking you agree to getting an email from time to time about new music I am releasing). Plus I could use the FB pixel and Google Analytics

    They ended up on my email list from that. From there they were in my funnel.

    Like I said, it was about 16 months to get an ROI (just a small one).

    1. Also by doing this i could see how many paid subs vs freeloaders and target emails differently, and could calculate the number of estimated plays a paid user did in a month, vs, a freeloader to determine the value of each. This allowed me to a/b test some emails with offers and found that yes, those that paid for spot also paid for CDs, and other things. usually about 2:1 compared to the freeloaders. So for the freeloaders, I had longer email campaigns to develop more relationship before hitting them up for a sale.

      1. That’s fantastic! I’m always scared of asking for too many permissions, but in your case it clearly paid off.

        Feature.fm provides that sort of granular data for presaves, and I’ve seen others use ToneDen gating to force playlist submitters to follow their playlists.

        The clickthrough rates for my two landing pages are currently 88% for the profile and 83% for the playlist. If I did the same thing and saw the same 60% drop, my ad costs would double.

        Whether or not it would be worth it is anyone’s guess!

  2. Wow. I’m still learning Indepreneur stuff and the jargon is still a little above me; But I’m impressed with your dedication for listeners on Spotify.

    1. Thanks Jonathan! I’m happy to clarify any terms you’d like. I’m so deep into this stuff that I don’t have a good sense of what’s jargon and what’s not.

  3. Brian,
    Nice article. Too bad indeprenuer keeps shutting your post down, over there. I think an open discussion would be better.

    I have questions 🙂

    If you’re spending $600 on ads, and getting back $300, is that balance changing? In other words, did it start at $600 with $100 back, and build from there, or is it a stable $1 back for each $2 spend?

    Is it your feeling that eventually the royalties will overtake the ad spend?

    If you stopped ads tomorrow, you’d still get some streams and royalty payments. What are your thoughts, or do you have any data on how long, or if the royalties would eventually pay back the ad costs?

    And finally, I’m having difficulty understanding where most of your royalties are coming from. You say 15% from algorithmic playlists. At $.30/follower, you’d have to get maybe a lot of streams from each follower to generate the other 85% of royalties. I don’t understand playlists well. Are there non algorithmic playlists that are generating revenue, or is the 85% from your followers?

    If there are non algorithmic playlists, what’s the revenue split between those playlists, and your followers?

    1. Yeah, it’s too bad they disabled comments. I understand where they’re coming from though.

      To answer your questions, I don’t think too hard about ROI. The streams directly generated from the ads are a very small percentage of my overall streams, and there’s no way to say how much they contribute to my algorithmic streams.

      If I stopped ads tomorrow, I suspect my overall numbers would drop a little bit. I’d see the biggest hit on Juggernaut, the song I’m using to promote both the playlist and profile. My playlist streams would definitely slow down though, and I genuinely want to help out the other artists on the playlist, so I wouldn’t do that to them.

      After promoting Juggernaut for so long, it gets 42% of the 15K streams per 4 weeks from “listener’s own playlists and library” and a ton of Shazams. Let’s see… 1846 total, 337 in the past 4 weeks.

      It also gets a ton of visitors to YouTube saying they heard the song on Instagram. The physics of the whole operation is just too much to take in.

  4. How do you get to that “Countries comparison” screen in Spotify for Artists? Can’t find it for the life of me.

    1. It’s only on individual songs. Go to Music ➤ Songs ➤ Song Name and it’s about halfway down. I just added a caption to one of the screenshots to clarify!

  5. Great post, as usual. As you point out this notion of losing $300/month (net of streaming revenue) on these facebook ad strategies it raises two related questions just for perspective: (1) What is the average net gain/loss that you see on your Color Theory efforts in a typical month [so, total revenue (Streaming + CD Sales + Merch sales + Patreon donations) less total expenses (Marketing + CD COGS + Merch COGS + Shipping costs + Misc)]? and (2) How many hours per month do you spend on all of the various aspects of Color Theory (aside from recording)? I’m not saying it’s not worth it – clearly, it is to you – but I’m curious as to how much time you have to dedicate to these various strategies on a monthy basis. It seems like a lot, but maybe I’m wrong.

    1. Let me just send you my tax return! 😉

      Kidding aside, there’s too much variation to even forecast my income. I’ve tried tracking my hours, but many activities don’t fit neatly into one category.

      The way I see it, I’ve got three main professional pursuits: my music, my mastering business, and this blog. My music feeds the other two, so regardless of whether or not it’s profitable on its own, it’s necessary. I’ve pretty much always had my fill of mastering and production work, without ever advertising.

      For tax purposes, all of my musical endeavors are lumped into one business, which has been profitable for as long as I can remember.

      1. Well, I’m certainly glad that your musical endeavors as a whole have been profitable given that mastering is your actual vocation – it would be problematic if that weren’t the case! Have you thought about doing more on the production front? I’d think the margins are better there (than mastering)… but I don’t know.

        1. I just finished mixing and mastering a full album, including editing all the vocals. I would absolutely make more money, at least in the short term, producing and mastering full-time.

          But hey, if I wanted to make money, I wouldn’t have gone into music in the first place!

  6. Great article. I have made an ad campaign following your advices. So far satisfied with the result. When trying to see location where my FB ad has been seen I get a lot of cities. Is there a way too see this location statistic as countries instead?

      1. Thanks, Brian. Got it.

        One more thing. I have been running a campaign for 4 days now and it’s gets ok amount of clicks. I also got a lot of messages indicating that a lot of people took it the whole way and actually heard the song. However according to the Spotify statistics there has been zero listeners from these countries. Any idea what the delay is from Spotify from the play until it’s registered at my artist profile?

        1. You should see the Spotify data the next day around noon Eastern time when it updates. Remember to look under Music ➤ Songs ➤ [Song Name] and scroll down to Country Comparisons.

          So you’re getting messages from listeners on Facebook and/or IG? Lucky you! I’ve never had that happen.

  7. Hi Brian, this is a great post, thank you so much for sharing all this information! I am also using a Toneden landing page and I want to run an ad that links direct to my spotify profile only, no other platforms for this particular ad. When setting up the smart link in Toneden, would you recommend checking the ‘Redirect all visitors’ box – ‘Skip landing page and send directly to your destination url’? I’m assuming this is sort of the same as using a url genius-type link but with the added bonus of being able to track the traffic via the landing page. I have done the custom domain and conversion event set up for my Toneden landing pages. Would I also need to optimise for Spotify links as you have in yours? Thank you, Happy Easter!

    1. Happy Easter Lucinda!

      I wouldn’t recommend skipping the landing page, because you risk optimizing for accidental clicks. A clickthrough from the landing page shows a higher degree of intent, and that’s what you’re looking for. Of course, your cost per conversion will be higher, but I’d guess your cost per follower will be lower over time.

      So I’d keep the landing page, even if Spotify is the only option. In that case, I’m not sure you need to define the event parameter in the custom conversion, but I would anyway just to be on the safe side.

  8. Hey man! So I keep seeing you a an a bunch of other people (a lot involved in John Gold’s Spotify growth engine, which I’ve also signed up for) try feed placements for Spotify traffic.

    I’ve tried it countless times over the years but it never ever leads to actual listeners. Cheap clicks every time, but almost nobody actually clicks through.

    Stories, I can usually achieve followers for about $1 each. anytime I’ve seen people beat that lately, it’s been feed placements. But I cannot feel the life of me get anything even close to $1/follow that way.

    Does it take a while?

    I’ve tried:
    – Turning on the wifi only setting
    – Conversion campaigns running for the second click on the bridge page
    – feed placements only
    – automatic placements
    – green light only countries
    – international countries.

    Always cheap clicks. Almost no listeners. Any idea why? I’ve even tried multiple platform and bridge pages.

    1. Hey Chandler!

      If you set it up the way I do, with separate videos for feeds and stories in the same ad, Facebook will optimize for whichever is cheaper. If that turns out to be stories, so be it!

      By the same token, you can test all those other things in one go, other than the wifi only setting, which hasn’t made a difference for me.

      I’d try a conversion campaign for clicks from the landing page, including all Spotify countries (except maybe my losers: Ukraine, Indonesia, Kazakhstan, Algeria, Tunisia, Morocco), with the placements I use (no Audience Network).

      If you’re seeing crazy cheap conversions, like less than $0.05, and no plays, maybe you’ve been optimizing for bots. I had that happen to me and I had to basically start over, tossing the lookalike I’d built.

      In that case, I’d narrow to something you know is safe (maybe just Brazil), and dial in the targeting – no lookalikes or detailed targeting expansion. Once you can see it’s working, expand little by little until you’ve had enough genuine conversions that you can build a lookalike from it.

      Let us know what you find out!

  9. Brian,

    I’m confused as to how you’re converting these clicks into followers. Are you using a download gate where they have to follow your profile/playlist to listen/download or are you just giving them the links to the song/playlist and hoping they’ll follow?

    Best,
    Tieran

    1. Ultimately I’m just hoping they’ll follow, but two factors increase my odds:

      1. I’m directing to my profile instead of a release, so the follow button will be front and center
      2. My landing page description is “Follow for new songs!”

  10. Brian, this is an amazing post. Thank you so so much for it!

    I am running ads on FB now and trying to learn how they work, so I will use this information… hopefully I can come back with some nice feedback!

    Just a quick question: are you still using the ToneDen landing page with all the links to your music stores? That is, you clarified that you “collected” the clicks for a lookalike audience from this landing page, but in your current campaign are you using that one or a new landing page that only directs to Spotify?

    All the best!

    P.S.: Yesterday I opened a ToneDen account using your referral link. Thanks for that as well!

    1. I am indeed still using a ToneDen landing page! The page lists seven services including YouTube and SoundCloud, but it only counts as a conversion when someone clicks on Spotify. The lookalike audience is created from those Spotify clickers. Make sense? I’m happy to clarify!

      Thanks for using my referral link!

  11. Thanks for the updated post on this. I had recently used your previous article on this subject and have had my ad account disabled for the 1st and now 2nd time. Have you pinpointed anything from your previous article that might be causing this? I was thinking that it’s possibly one of the following:

    1) The Genius URL deep linking
    2) The phone scroll style ad (could possibly be considered misleading)

    Any thoughts on this since these were both things that you did within the first article?

    1. So sorry to hear that Jay!

      I highly doubt either of those are the issue. If the scrolling ad were misleading, they would’ve rejected it and cited that reason.

      I’m guessing they just disabled the account like they did with my friend, with no clear reason or guidance. So frustrating!

      My advice is to politely but firmly state that you don’t think you violated any policies, and to ask that your account be reinstated, through whatever channels are available.

      If you can trace the reasoning back to anything I advocate here, please let me know and I’ll update any relevant articles!

      1. Just an update. I did get my account re-activated. Of course there was no explanation. I did go through their little course that walks you through a bunch of possible term violations and did did learn that you’re not allowed to have fake calls to action on your ad. Since there are play buttons on the Spotify scroll ad, that could potentially violate those terms. Perhaps they have heat maps on these ads that are auto-flagging odd clicking/tapping behavior. Who knows?

        1. Hallelujah!

          The fake CTA seems like a stretch, but your guess is as good as mine.

          Amazing that we can pay Facebook hundreds of dollars a month and get zero support when we run into a problem. At this point I feel like I should have a dedicated hotline.

Leave a Reply

Your email address will not be published. Required fields are marked *