How to test ad hooks before you spend

The expensive way to test a hook is to film five videos, run them as five ad sets, and let a platform spend your money finding out that three of them were never going to work.

There are cheaper ways to get most of that information. None of them replaces a live test — but they all narrow the field before it costs anything, and the narrowing is where the money is.

Why choosing is harder than writing

Reading your own hooks back is close to useless, and it is worth understanding why before trusting any method.

You read them in order, so the first one anchors your judgement. You read them in your own voice, which is not the voice a viewer will hear. You read them knowing what the product is, which no scrolling stranger does. And you read them slowly, in a document, in a quiet room, having already decided which one you like.

Every method below is really a trick for removing one of those four advantages you have and a viewer does not.

1. Forced pairwise comparison

Put two hooks side by side and pick one. Then two more. Then the winners against each other. Four rounds gets you from eight candidates to one, and each decision takes about two seconds.

What it removes: order effects and slow deliberation. A pair is a decision; a list is a review.

What it tells you: which line has more pull for a reader who has to choose right now. That correlates well enough with scroll-stopping to be worth doing.

What it cannot tell you: whether the winner is true, or whether it survives being filmed. A hook can win on the page and collapse when a human has to say it.

Cost: minutes. This is the one to do first, every time, and it is what Hook Battle automates — it drafts eight hooks from your product page and runs the four rounds for you.

2. Show it to five people who do not work for you

Send the five openers, one per message, to five people who fit the buyer and do not know the product. Ask one question: what do you think this is selling?

What it removes: your knowledge of the product.

What it tells you: whether the hook is comprehensible. You will find at least one line that everybody misreads, which is information you cannot get any other way.

What it cannot tell you: appeal. People are polite, and a friend’s “that one sounds good” is worth nothing.

Cost: an hour of goodwill. Do it once per product, not once per campaign.

3. The three-second mute test

Film the openers roughly — phone, no lighting, no retakes. Play the first three seconds of each with the sound off to someone who has not seen them, and ask what the video is about.

What it removes: your voice and your pacing. A line that only works because of how you say it is a line that will not survive a creator reading it.

What it tells you: whether the visual and the line agree, and whether the opener survives muted autoplay — which is how a large share of views begin.

What it cannot tell you: anything about the rest of the video.

Cost: twenty minutes of filming. Worth it before you book anything.

4. Post it organically first

Put the two or three survivors out as ordinary posts on your own account before you put money behind any of them.

What it removes: all of your bias, replaced by real feed conditions.

What it tells you: the only signal that matters — do strangers watch past the opening. Watch retention in the first three seconds, not likes, not comments, not shares. A hook can be widely liked and still fail to hold anyone.

What it cannot tell you: how it behaves in a paid placement, with a cold audience and a different delivery pattern. Organic reach is not paid reach.

Cost: a few days and some patience.

5. Then, and only then, spend

By the time you pay for reach, you should be testing two openers you already believe in, not five you drafted last Tuesday.

Two rules for the paid test itself:

Change only the hook. Same body, same footage, same CTA, same landing page. If two ads differ in three places, a winner tells you nothing about which of the three mattered.

Give it enough to be a result. A difference that appears in the first few hundred impressions is usually noise. If you cannot fund a test to the point where you would act on the outcome, do not run it — spend the money on making one better ad instead.

The order that works

1. Write six to eight genuinely different openers — one per objection, not six rewrites of your favourite.
2. Run them head to head until one is left.
3. Mute-test the top two on rough footage.
4. Post those two organically.
5. Put money behind the one that holds attention.

Most of that costs nothing but an afternoon. The step people skip is the first — writing more than one — and every method here is worthless without it. The method for producing several and the formulas to draft from are the place to start.

Related

• How to write a hook for a product video — producing the candidates
• The first 3 seconds of a TikTok ad — what the mute test is actually checking
• TikTok hook ideas for product videos — thirty starting points