Why test the parts
You’d think the chorus sells the song. Often it doesn’t.
“When my album came out, I had about a thousand monthly listeners and thirteen songs I loved. I started with Mirror, the first song, and paid 50 to 60 cents for every click to Spotify or Apple Music. I spent a lot of money that way, because making new videos in CapCut was so painful.”
Kai Williams, music artist and creator of Clips
Kai’s next song, Bad to You, has a long intro, then a verse, a chorus and a bridge. He ran the chorus for weeks at $0.63 a click, then made a video for each part. The intro came in at $0.21, a third of the chorus. The part he would have skipped was the cheapest way to find listeners.
Brian Hazard of Passive Promotion runs ads for his own music and publishes the results. In one campaign, the winner was “this little snippet of the last verse going into the coda”, which outperformed everything else in the test. Hazard’s advice now is to make videos that “showcase at least three different parts of your song”.
A stranger scrolling Instagram gives your ad about a second. The part of the song playing in that second decides whether they stay. You can’t know in advance which part that is, but a week of ads can tell you.
What you need
- The song, mixed and mastered.
- Short vertical videos, one per part (step 1).
- A smart link with a pixel, such as Hypeddit, Feature.fm or DistroKid’s HyperFollow. Spotify won’t let you put a Meta pixel on your artist page, so the smart link sits in between. People land on it and click through to Spotify, and the click gets counted.
- A Meta ad account, set up in Meta Business Suite and linked to your Instagram.
- $5 to $15 a day for about a week.
Step 1: Make one video per part
Pick at least three parts: the first verse, the pre-chorus, the chorus, the bridge, the outro. Make one short video for each.
- 1
Vertical.
9:16, Meta’s spec for Reels.
- 2
Short.
6 to 15 seconds. Meta says shorter videos in that range work better.
- 3
Start on the hit.
Not on the part’s first bar. Meta’s advice is to get your key message into the first 3 seconds and to use motion in the first frame.
- 4
Change only the music.
Use the same backdrop, look and text style on every video, so the only difference is the music. Then the test tells you which part won, not which background did.
- 5
Mind the edges.
For Reels, Meta suggests leaving about 14% at the top, 35% at the bottom and 6% on each side free of text. Drop a frame into the free safe zone checker to see what the app covers (there’s one for Reels, Stories, TikTok and YouTube Shorts).
- 6
Show the lyrics.
Meta recommends sound, but people often scroll with it off, and words on screen still tell them what they’re hearing.
Step 2: Set up the campaign
- 1
Pick the objective.
Set the objective to Engagement, with Website as where conversions happen. Choose the pixel event your smart link sends for a click through to Spotify. This tells Meta to find people who click through, not people who just tap.
- 2
No tracking?
If you can’t track conversions, optimize for landing page views instead. A landing page view counts only when your page actually loaded, so it’s closer to real interest than a link click.
- 3
One ad set.
Broad targeting. Pick the countries you care about and leave interests open. As Hazard puts it, “the creative is the targeting”: the song itself finds its listeners.
- 4
One ad per part.
Give each its own video, so you can compare them. Meta suggests up to 10 creatives in an ad set.
- 5
Instagram only.
Pick it for placements. Hazard found that adding Facebook made the cost per result look cheaper with no gain in Spotify for Artists.
- 6
Budget.
$5 to $15 a day. Hypeddit suggests $5–10; Hazard tests at about $15.
Step 3: Let it run, then read the numbers
Give it at least 4 days, ideally a week, before you judge. Meta suggests at least 4 days for creative tests and 7 for its A/B test tool. Switching ads on and off every day ruins the test.
| Number | What it tells you | How to read it |
|---|---|---|
| Cost per result (Spotify click) | What a listener costs you | The main number. Lower wins |
| Hook rate (3-second plays ÷ impressions) | Whether the first second stops the scroll | 20–40% is solid, according to Motion’s ad benchmarks |
| Hold rate (ThruPlays ÷ 3-second plays) | Whether the part keeps people listening | Higher means the part holds |
| Saves and followers in Spotify for Artists | Whether the clicks turn into fans | Check it after the test week |
A part with a great hook rate but expensive clicks stops people but doesn’t make them curious. A part with fewer views but cheap clicks is the one to fund.

Step 4: Put the money behind the winner
- 1
Keep the best.
Keep one or two parts and turn the rest off.
- 2
Refresh the winner.
Make new videos of the winning part with a different backdrop, first second or hook line (the short text on screen that sets the scene). Meta says ads that are too similar may not be treated as different, so make the new ones clearly new.
- 3
Raise slowly.
Raise the budget gradually, and keep watching saves and followers in Spotify for Artists. They tell you whether the clicks are real fans.
- 4
Test again.
The winner on one song tells you little about the next release.
Instagram, Facebook, TikTok or Spotify ads?
| Where | Good for | The catch |
|---|---|---|
| Instagram (Meta) | New listeners; testing every part cheaply | Needs a smart link to count Spotify clicks |
| Facebook (Meta) | Same ads, same campaign as Instagram | Cheaper-looking results, no more listeners (Hazard) |
| TikTok | Boosting a post that already works | $20 a day minimum per ad group |
| Spotify (Marquee, Showcase) | People already on Spotify | Needs 1,000 streams in 28 days in a market, and $100 (Marquee also 5,000 active listeners) |
On TikTok, post each part as a normal TikTok first, see which one people watch to the end, then boost that post as a Spark Ad. A Spark Ad keeps the likes and comments on your original post. TikTok’s minimum budget is $20 a day per ad group, so testing many parts as separate ads costs more there than on Instagram.
On Spotify, Marquee shows a new release as a full-screen pop-up and Showcase as a banner on Home. Both are bought in Spotify for Artists and charged per click, with saves counted as clicks. They work best once you have listeners in a market; Instagram is where you find them first. See Spotify ads for artists.
What the listeners from your ads do next, the saves, follows and repeat plays, is what Spotify notices. The guides to the Spotify algorithm and Release Radar, the popularity score and monthly listeners show how. If you’re weighing a paid playlist pitch instead, see Spotify promotion services compared.
Make the videos in one go
The slow part is step 1, cutting the same song five or fifteen ways, on the beat, with the lyrics in time. That’s an evening in a video editor, so the temptation is to make two videos and hope.
“In CapCut, the automatic lyrics came out wrong, and fixing them was buggy. Final Cut was worse. Every new hook line or backdrop meant opening the editor again. I was losing whole evenings I could have spent making music.”
Kai Williams, music artist and creator of Clips
Clips does it in a few minutes. Drop in the song and your footage, and it finds the parts. It cuts one vertical video for each on the beat and syncs your lyrics word by word. The look and layout are the same on every part, so your test is fair. The files come out named by part, ready to upload as separate ads.
Sources
10
- What's working with Meta ads for music and Facebook Spotify ads best practices, Brian Hazard, Passive Promotion
- Instagram Reels ad specs, Meta
- Video ad best practices, Meta Business Help Center
- About A/B testing and test length, Meta Business Help Center
- Creative differentiation, Meta
- Meta Ads conversion campaigns, Feature.fm
- How to get more monthly listeners on Spotify, Hypeddit
- Thumbstop rate and hold rate, Motion
- Spark Ads and budget, TikTok Ads
- Getting started with display campaigns and Marquee and Showcase budgeting, Spotify for Artists



