By the time a song is finished, you've heard it hundreds of times and looked at the cover for weeks. You can no longer tell what a stranger will notice in the first second. A small test before release gives you a second pair of eyes and ears. It doesn't need a budget or special tools, but it does need a plan, because a careless test is worse than no test.
What is worth testing
Test things that a listener decides in a moment:
- Cover art: two or three options at most. On a streaming app, the cover is small, and it is often the reason someone presses play or scrolls past.
- The hook: the part you would put in a short video or a story. Which fifteen seconds make a person want more?
- The title: only if you're really torn between two. A title is hard to change later.
Don't test the whole song or the mix this way. Those need real listeners and time, not a quick poll.
Ways to run a quick test
You have three simple options, from the cheapest to the most structured.
- Stories or a poll. Post two cover options side by side and ask which one makes people want to tap. For hooks, post two short clips on different days or to different halves of your audience. Keep the wording identical, so the only difference is the thing you're testing.
- Close friends or a small group chat. A list of ten to twenty people, some of whom are not your best friends, works well. Ask them separate questions rather than "do you like it?": which one would you tap first, what kind of music does this look like, which part do you remember?
- A small ad. If you already advertise your music, show two versions to a similar audience at the same time, with the same budget and the same duration. Compare how many people stop, watch to the end or click.
Whichever method you pick, decide in advance what counts as a win. If you only look at the numbers after you see them, you'll choose the ones that agree with you.
Testing a cover
A cover has a few jobs, and each can be tested quickly.
- Size test. Shrink it to the size of a thumbnail on your phone. Can you still read the title or recognise the image?
- Genre test. Show it for two seconds and ask what kind of music it would be. If the answer doesn't match your music, listeners who like your genre might skip it, and listeners who don't might click and leave.
- Contrast test. Look at it in black and white and in a dark mode screen. A cover that depends on small details disappears in a feed.
- Text test. Check that any text is large, spelled correctly and not too close to the edge, and that the artist and track names don't clash with the platform's own text.
Whichever cover wins, export it as a JPG at exactly 3000 × 3000 pixels, which is the size the CULT portal asks for, and check that you own or have permission to use every image and font in it.
Testing a hook
For a hook, the first three seconds matter most. Cut two clips of the same length, starting at different places: one on the chorus line, one on the verse. Watch what people do. Do they stay to the end? Do they send it to a friend? Do they ask for the full song?
A second test is memory. Play the clip once for a friend and ask them the next day to hum or describe what they remember. If they can't recall anything, the hook might be too complex or not placed where you thought.
How to read the reactions
- Treat a few likes as noise. A handful of answers is not proof, so look for a clear difference, not a tiny one.
- Give more weight to actions than to opinions: shares, saves, replies, full views.
- If the results are split evenly, both options probably work, so choose the one you prefer.
- Remember who answered. Your followers already know you, so their answers tell you about your fans, not about a new listener.
Don't over-test
Testing is a tool for avoiding a clear mistake. It isn't a way to remove all risk. One or two rounds are enough. If you test eight covers across six weeks, you'll delay the release and end up with the most average option, which usually means the least memorable.
If feedback contradicts your artistic idea but the idea matters to you, it's fine to keep it. Use the test to learn what people notice, then decide on purpose.