# How to Test a Song Before You Release It

> Four tests you can run on an unreleased song, what each one tells you, and the feedback question that produces useful answers.

[All guides](/blog)

# How to Test a Song Before You Release It

Published 28 August 2026, 18:00 UTC · updated 12 August 2026, 21:30 UTC · by Loopin

You can find out most of what a release will do before you release it. None of the four tests involve asking anyone whether they like the song.

## Test one: the eight-second clip

Post the hook as a clip three weeks out with no mention of a release. Retention at three seconds tells you whether the section grabs strangers, which is the single most predictive number available before release.

Run it on two or three different sections. If none hold 30%, you have found something important while there is still time to change the single.

## Test two: the skip point

Send the full track to ten people and ask one question: "where did you get bored?" Not whether they liked it — where their attention went. People answer that honestly because it is not a judgement.

If six of them name the same bar, that is an arrangement note and it is worth acting on. If the answers scatter, the song is fine and you can stop worrying.

## Test three: the next-day recall

Play it once for someone, then message them the next day and ask them to hum what they remember. Whatever comes back is your actual hook, and it is frequently not the part you thought.

This is the cheapest way to find out whether a song is memorable, which is a different property from whether it is good and matters more for a single.

## Test four: the small-speaker test

Play it on a phone speaker at low volume in a room with background noise. Everything below 500 Hz vanishes. If the song falls apart, it is leaning on the low end more than real listening allows.

This also catches masters that only work loud. [Master at a sensible target](/mastering) and check it here rather than on monitors, because this is closer to how it will actually be heard.

## What not to ask

"What do you think?" produces politeness. "Do you like it?" produces politeness. "Is it good?" produces politeness. Every useful question is about behaviour: where did you stop, what do you remember, would you send it to anyone.

And do not ask other musicians exclusively. They listen for craft; your audience listens for feeling, and the two give different answers.

## Frequently asked questions

### How do I test a song before releasing it?

Post the hook as a clip three weeks out and read retention at three seconds; ask ten people where they got bored; ask someone the next day to hum what they remember; and play it on a phone speaker in a noisy room.

### What should I ask when getting feedback on a song?

Questions about behaviour, not opinion. "Where did you get bored?" and "what do you remember?" produce honest answers. "Do you like it?" produces politeness.

### Who should I get feedback from?

Not only other musicians — they listen for craft while your audience listens for feeling. Include people who just listen to music.

## Related reads

- [How to pick which 8 seconds to post](/blog/how-to-pick-the-clip-that-goes-viral)
- [How to give feedback on songs](/blog/how-to-give-feedback-on-songs)
- [The teaser schedule that actually builds](/blog/the-teaser-schedule-that-works)

---

Source: https://onloopin.com/blog/how-to-test-a-song-before-you-release-it