To test a social media hook, publish the same post twice with different opening lines, at the same hour, on comparable days, and compare each to your own median. Change nothing else. Then repeat, because one result is rarely proof. No hook reliably doubles reach, but a clean test will show you which openings stop your audience and which do not.
TLDR: Write two hooks for one idea, publish them at the same hour on two comparable days, and compare the engagement rate to your median. A gap of 20% or more across several rounds is a signal. A single gap is noise.
What a hook is and why it needs its own test
The hook is the first line of a caption or the first two to three seconds of a video. It is the only part most people see before they decide to stop or scroll on.
That makes it a good thing to test: it is small, quick to change and it comes before everything else in the post. But it is also easy to test badly, because the hour, the topic and the format all move the result more than the wording does. In Mydrop's September 2026 data, posting between midnight and 6 am cost Instagram posts 36% of their engagement. A hook test run at the wrong hour tells you about the hour.
The clean test in four steps
1. Pick one idea and write two hooks. Same post, same format, same visual. Only the first line changes. For example:
- Version A (result): "We cut client approvals from 3 days to 1."
- Version B (question): "Still waiting 3 days for a client to say yes?"
2. Fix everything else. Same hour, same format, similar weekday, no paid boost on either. Do not post both on the same day on the same account, because they will compete for the same followers.
3. Publish and wait 48 hours. Early numbers move a lot. Read the engagement rate and reach after two days, when most of the post's life has happened.
4. Compare each to your median. Divide each post's engagement rate by your median over the last 30 posts. A post at 1.3 beat your normal. A post at 0.7 missed it. Compare A and B by that score, not the raw numbers.
If you have a tool that tests variants natively (Instagram offers Trial Reels, which show a Reel to non-followers first), use it for video hooks. Otherwise the sequential method above is enough.
How to read the result without fooling yourself
| What happened | What it means | What to do |
|---|---|---|
| A scores 1.3, B scores 0.9 | A may be better | Run it again with a new idea |
| A and B both near 1.0 | The hook was not the lever | Test the format or the hour instead |
| Same winner in three rounds | A real pattern | Add that hook type to your library |
| Winner flips each round | Noise | Stop testing and look at topic |
The honest rule: a hook type has to win at least three rounds before you build a content plan around it. In our data a normal Instagram post earns between 1.1% and 6.7%, a range wide enough to hide a small gap.
Hook ideas worth testing first
Test big differences, not small word swaps. "Great tips for Instagram" versus "Top tips for Instagram" will tell you nothing. These pairs differ enough to show something:
- Result versus question: lead with the outcome or ask the problem.
- Number versus no number: "5 mistakes" versus "the mistake".
- Your voice versus a customer's: "I" versus a quote from a client.
- Short versus long: 6 words versus 15.
Keep a note of the winners in one list. Once you have run a few rounds, build a weekly hook log so the results add up instead of disappearing.
Make the testing easy to repeat
Testing falls apart when it depends on your memory. Put both versions in the content calendar on the same weekday and hour, so the setup stays identical. Mydrop's performance view compares each post to the account's own median, which gives you the 1.3 or 0.7 score without a spreadsheet. AI captions can draft five hook variants in a minute so you spend your time choosing, not typing.
Write a second hook for this week's post
Pick the post you are about to publish this week. Write a second hook for it, schedule the two versions at the same hour on two comparable days, and check both after 48 hours. If you want the scoring done for you, start free with Mydrop.
Google review
Trustpilot review
Product Hunt comment
