Do question hooks beat statement hooks on short videos?
A planned test on my own short videos, comparing openings that ask a question with openings that make a claim.
By Jayaprakash DasMarketer, researcher and speaker
The LabEntry #002
Details
- Status
- Planned
- Evidence
- Illustrative
- Format
- A/B-test diary
- Topic
- How growth happens
- Status
- Planned
- Evidence
- Illustrative
- Reading time
- 2 min read
- First published
The question
Do short videos that open with a question keep people watching longer than videos that open with a statement?
Hypothesis
Videos that open with a question will hold viewers past the first three seconds more often than videos that open with a statement, because a question opens a loop the viewer wants closed.
This is a sample entry that shows how an experiment looks. It will be replaced by a real one.
Setup
Most advice on hooks is confident and untested. My own account is small, so this is a chance to check one piece of that advice on real viewers rather than repeat it. If the hypothesis is wrong, I expect the difference to be small or to vary by topic.
- Ten pairs of videos on the same topics. Each pair shares a script after the first line; only the opening changes.
- One version opens with a question (“Why do discounts sometimes reduce sales?”), the other with a statement (“Discounts can reduce sales.”).
- Each pair is posted a week apart, on the same weekday and at the same time, alternating which version goes first.
- Measured from the platform’s own analytics: three-second hold, average watch time, and saves and shares per 1,000 views.
Baseline
The baseline is the average of my last ten videos before the test starts, on the same three measures. It will be recorded here before the first pair is posted, so the comparison cannot be adjusted afterwards.
| Measure | Baseline | Question hook | Statement hook |
|---|---|---|---|
| Three-second hold | to be measured | to be measured | to be measured |
| Average watch time | to be measured | to be measured | to be measured |
| Saves and shares per 1,000 views | to be measured | to be measured | to be measured |
Findings
Nothing has been measured yet. When the test ends, this section will report the numbers for each pair, the overall difference and whether it was consistent across topics, including any pairs where the statement hook won.
Limitations
- Ten pairs is a small sample, and one account’s audience is not everyone’s.
- Posting a week apart means the algorithm, the news and the audience can all change between the two versions.
- Platform analytics are not fully transparent, and I cannot control who is shown each video.
Next experiment
If question hooks do better, the next test is which kind of question works: a curiosity question or a problem question. If there is no clear difference, the next test moves to what follows the hook, since the opening may matter less than the advice suggests.