Guide
How to test two video openings (hook tests)
Make two versions of the same film that differ only in the first few seconds. Show each to a similar audience at the same time and budget, decide the metric before you start, and read the result only once enough people have watched. Small differences need many views, so test big differences first.
A product video is often judged in its first few seconds. Instead of arguing about which opening is better, you can test two. This guide shows a simple method. It is our own guidance, not a study, and platforms change their tools and metrics, so check what your platform reports.
Step 1: decide what you are testing
Write one sentence: "I think a [problem-first] opening will keep more people watching than a [result-first] opening." One idea per test.
Good pairs to test:
- Problem first against result first
- A question against a plain statement of what the product does
- A close-up of one screen against a wide view of the whole product
Step 2: change only the opening
Keep everything after the first few seconds identical: the same music, the same product footage, the same ending. Then any difference in results belongs to the opening. Our guide to what makes a good hook lists what a strong opening shows.
Step 3: pick the metric first
Write the number down before you run anything.
| Question | A sensible metric |
|---|---|
| Does the opening stop the scroll? | Share of viewers still watching after the first few seconds |
| Does the film keep them? | Share who reach the end, or average watch time |
| Does it lead somewhere? | Clicks to your page or sign-ups from each version |
Step 4: split the audience evenly
- Show both versions to similar people at the same time, with the same budget and placement.
- If you run paid ads, use the platform's own split-test option so the audience is divided fairly. If you post organically, post at similar times on similar days and treat the result as a hint, not proof.
Step 5: wait for enough views, then stop
A few dozen views tell you nothing. As a rough rule from our own calculation, a big difference in opening retention can show up with a few hundred views each, and a small one needs thousands. If the two are close after a reasonable test, call it a tie and test something bolder next time.
Mistakes that spoil a test
- Stopping the moment one version leads. Early numbers swing wildly. Decide the number of views first.
- Testing a tiny change. A different word in the title rarely moves anything you can measure.
- Running one version on Monday and the other on Friday. Different days bring different audiences.
What a test can and cannot tell you
- It can tell you which opening holds attention better for that audience, on that platform, now.
- It cannot tell you whether the film will make sales. That also depends on your product, your offer and the page people land on.
A worked example with made-up numbers
Say you run two openings for a fictional app called ExampleApp, each shown to a similar audience. Opening A keeps 30 of every 100 viewers past the first few seconds and opening B keeps 40. That ten-point gap is large, so a few hundred views each could show it. If A kept 30 and B kept 32, you would need far more views, and the sensible move is to call it a tie and test a bolder change. The numbers here are invented to show the idea only.
How we build for testing
Our Launch Day package includes two openings to test, and Launch + Ads includes three openings in three shapes, so you can run this method without making new films. See the ads package or the launch film page.
Frequently asked questions
What should I change between the two openings?
Change only the first few seconds, and only one idea at a time: for example, a result-first opening against a problem-first opening. If you also change the music, the title or the ending, you will not know what caused the difference.
How many views do I need before I trust a result?
It depends on how big the difference is. As a rough rule from our own calculation, a large gap between two openings shows up with a few hundred views each, while a small gap needs thousands. If the numbers are close after a modest test, treat it as a tie and move on.
Which metric should I use?
Pick one before you start. For the opening, use the share of people who keep watching past the first few seconds on the platform you are using. For the whole film, use how many reach the end or take the next step. Choosing after you see the data makes any result look good.
Can a hook test guarantee more customers?
No. A test shows which opening holds attention better for your audience at that moment. Whether that turns into customers also depends on your product, your offer and your landing page. Use it to stop guessing about the opening, not as a promise of results.
Keep reading
What makes a good video hook for a SaaS product?
The first few seconds of a product video decide whether people keep watching. Here is what a strong hook shows, and a quick test to run on your own opening.
Video aspect ratios for a product launch
You usually need 16:9, 9:16, 1:1 and 4:5. Here is where each fits, how to lay them out instead of cropping, and why to check platform specs before exporting.
Want a film like this made for you?
Send a ten-minute brief. We send a one-page storyboard first, then a private preview. You pay only once you are happy with it.
Get a free storyboard