Creative metrics beyond ROAS: thumbstop, hold rate and click through as your diagnosis layer

ROAS tells you that an ad fails, not why. Thumbstop, hold rate and click through pinpoint where your creative leaks, so you iterate on purpose instead of guessing.

ROAS tells you whether an ad works, but not where it goes wrong. For that you need a diagnosis layer: thumbstop, hold rate and click through. Thumbstop measures whether your hook makes people stop scrolling, hold rate whether your story holds them, and click through whether your offer moves them. Read those three in sequence and you see exactly which part of your creative leaks, so you iterate on purpose instead of building a random new ad.

Why is ROAS alone not enough?

ROAS is an outcome metric. It compresses the whole journey from scroll to purchase into one number, and that is precisely its limitation. When an ad has a low ROAS, you know something is wrong, but not whether it is your hook, your story or your offer. You are then left choosing to remake the whole ad, and that is guessing. The diagnosis layer breaks that single outcome into three steps, so you know which link to replace.

The beauty is that each step tests a different part of the creative. So you never have to throw away a winning hook because the offer did not work, or bury a strong offer under a weak opening. You only repair what is broken, and that saves time, budget and frustration.

What exactly does thumbstop measure?

Thumbstop, also called the hook rate, measures what share of the people who see your ad stay with it in the first seconds instead of scrolling on. It is the test of your opening. A low thumbstop means your first seconds do not stop people: the wrong viewer, a slow start, or a visual that does not spark curiosity. It says nothing about your product or offer, because the viewer has not even seen those yet.

If your thumbstop is low, you work on the opening: a stronger visual hook, a more direct problem statement or a result you show straight away. All the other metrics only become relevant once people have actually stopped, so this is the first knob you turn.

What does hold rate tell you?

Hold rate measures whether your story keeps the viewers who stopped. Someone stops because of your hook, but do they drop off again after three seconds, or keep watching until you reach the offer? A good thumbstop with a low hold rate is a classic pattern: the opening pulls people in, but the story after it does not hold up. In that case the leak is in the middle of your ad, not in the start.

A strong hook with a weak hold rate is a promise your story fails to keep.

With a low hold rate you look at pace and structure. Does it take too long before anything happens? Does the story jump from the hook to something the viewer did not expect? Is it missing the proof that makes the hook's promise credible? This is the part you rewrite and re-edit, while you leave the hook that did work in place.

How do you read click through?

Click through measures whether the offer actually moves people. Your viewer stopped, kept watching, but do they also click through? A good thumbstop and hold rate with a low click through points to a weak or unclear offer, a missing call to action or a reason to act now that does not convince. The story did its job, but the final step is missing.

• Low thumbstop: rebuild your opening, because people do not even stop.
• Good thumbstop, low hold rate: rewrite the middle, because your story does not hold up.
• Good hold rate, low click through: strengthen your offer and call to action, because people do not move.
• Everything good but low ROAS: then the leak is further on, on your landing page or in your offer after the click.

How does this make your iterations more targeted?

The big advantage is that you stop guessing. Instead of remaking an entire ad for every disappointing ROAS, you read the three metrics and replace only the part that leaks. You combine a winning hook with a new middle, a strong story with a sharper offer. That way you build on what works and waste no production time on parts that were already good. That is exactly how we iterate creatives: diagnosis first, then targeted rebuilding.

With 15,000+ creatives built we have learned that the fastest route to a winner is not more volume, but better diagnosis. Anyone who knows where an ad leaks reaches the same win with half the tests. The metrics are not dashboard decoration, they are your tool to make the next version better.

Conclusion

ROAS says whether an ad works, thumbstop, hold rate and click through say where it fails. Read them in sequence and you know whether to rebuild your opening, your story or your offer. That way you iterate on purpose and reach winners with fewer wasted tests. Want to tackle a diagnosis layer that sharpens your iterations? Book a call and we will gladly look at it with you.