An artist sent me a screenshot of their Meta ad campaign, thrilled. Forty thousand impressions, an eight percent click-through rate, thousands of views. "It's working, right?" Their streams hadn't moved. It was not working, and the reason is that almost every number on that screen is a distraction from the one that isn't.
Here's the number that isn't: cost per click-out. What it costs you for a real person to press play on your music.
Why every other number lies to you
Impressions count eyeballs that scrolled past. Views count people who watched a video that was going to play automatically anyway. Click-through rate counts taps, and a lot of those taps are bots, misclicks, or people who bounced the instant your landing page loaded.
None of them count the only action that matters, which is a human choosing to go listen to your song. That action has a name in your ad dashboard once you set it up right: a click-out, or a conversion, or a landing-page click to Spotify. And it has a cost. That cost is your scoreboard. Everything else is the crowd noise.
Here's the mental model: you are not buying streams. You are buying a person's decision to press play, and then the music has to do the rest. If the ad is cheap at buying that decision, your ad is good. If the song keeps the people who pressed play, your song is good. Two different jobs, two different numbers, and confusing them is why artists throw money at a good song with a bad ad or a good ad with a song nobody saves.
The benchmarks
Rough numbers for indie budgets, and treat them as a floor to graduate from, not a ceiling to celebrate:
Around 80 cents per click-out is a normal first campaign. You're learning, the audience is cold, this is fine.
Under 50 cents is a strong result. Your creative and targeting are landing.
Under 30 cents, even under 20 is what experienced people hit after they've tested a lot of creative. It's absolutely reachable, and if you're consistently under 50, your new goal isn't to feel good, it's to beat 30.
Well above two dollars means something is broken. Fix the creative or the audience before you add a single dollar. Money never rescues a losing ad, it just loses faster.
And the most important benchmark isn't on any chart: from your second campaign onward, the number to beat is your own best. Generic bands are for beginners. Once you have a personal baseline, that's your target, and it should keep dropping.
A click-out is not a stream and not a save
This trips people up, so let me be exact. Expect fewer streams than click-outs. Spotify only counts a stream after 30 seconds from a logged-in listener, and some people who click never press play, or press play and skip. That drop-off is normal, not a broken campaign.
So judge the ad by cost per click-out. Judge the song by saves in Spotify for Artists. If the ad is cheap and the saves are still flat, you have a song problem, not an ad problem, and no budget fixes a song problem. If the ad is expensive but the people who do click out are saving like crazy, you have a creative problem in front of a song people love, which is the good problem, because creative is fixable.
The friction that looks like a leak
One thing artists try to optimize away, and shouldn't: the extra step. They want the ad to go straight to Spotify to "reduce friction."
Don't. Send the ad to a landing page, your smart link, and let people choose their platform there. The person who saw your ad, stopped scrolling, tapped listen, landed on your link, and picked Spotify just made four or five conscious decisions. By the time they press play, they're invested. That's exactly why ad-driven listeners save and follow at far higher rates than people who got fed your song on a playlist. The clicks aren't friction bleeding away streams. They're a filter turning scrollers into fans, and as a bonus, that extra page is where your pixel lives so the ad can actually optimize.
When to kill it
Campaigns that stay expensive after the learning window get one honest process, not more money:
First, swap the audience. It's the cheapest test, and often it's just the wrong people seeing a fine ad.
Still expensive? New creative, and I mean a genuinely different angle, not a recolor of the loser.
Both failed? Then the hard truth: this song isn't connecting with cold audiences right now, and the budget is better spent on your next release. That's not a defeat. That's the campaign doing its second job, which was always to test whether the song had the pull to justify more money. You cannot force a song to connect, and the numbers don't lie.
One more expectation to set before you spend a dollar: the streams often keep coming after the campaign ends. Practitioners consistently see algorithmic pickup lag about two weeks past the last ad dollar, so judge a campaign two to three weeks after you stop, not the day you stop.
The one-sentence version
Ignore impressions, views, and CTR. Watch cost per click-out to judge the ad, watch saves to judge the song, send traffic to a landing page, and beat your own best number every time out. That's the whole discipline, and it's a core part of what the ads coaching inside Musuni walks artists through, but you can start reading the right number tonight. Most artists never do, which is exactly why the ones who do pull ahead.