Measure creative on purchases, not on the ad.
Hook rate, hold rate and thumbstop are diagnostics, not outcomes. A note on which numbers should decide what gets made next.
This piece is now part of a longer guide: Landing page vs product page for paid ads: the full guide.
Every creative team has a dashboard full of numbers that feel like performance and are not. Hook rate. Three-second views. Hold rate. Thumbstop ratio. They are useful, and none of them should decide what you make next.
The difference between a diagnostic and an outcome
A diagnostic tells you where inside an ad something went wrong. An outcome tells you whether the ad was worth making.
Hook rate is a diagnostic. It tells you the opening did its job. It cannot tell you whether the people it held were people who would ever buy. We have all shipped a creative with a beautiful hook rate and nothing behind it, and we have all had the uncomfortable meeting afterwards.
Why teams drift toward the wrong number
Because the wrong number moves faster. Purchases are slow, noisy and often sit behind someone else's attribution setup. Hook rate is available this afternoon.
So teams optimise the metric that answers quickly, and three months later the library is full of ads that are very good at stopping the scroll and mediocre at everything after it.
What we report and what we sign to
We separate the two deliberately.
What we report is the scoreboard: completed purchases per thousand impressions. It is the number that tells a client whether the whole system is working, and it does not care which part of it we own.
What we sign to is narrower and more honest: holdout-measured creative lift, and share of spend captured. A creative set claiming a meaningful share of total campaign spend is a strong signal the auction is finding it, and that is a fact about the market rather than a claim about ourselves.
The rule we use internally
A number gets to decide what we make next only if it survives the question: if this went up and revenue stayed flat, would we still be happy?
Hook rate fails that test. Hold rate fails it. Share of spend mostly passes. Purchases per thousand impressions passes outright.
Where diagnostics still earn their place
Inside the ad. Once you know a creative worked, the diagnostics are how you find out which part did the work, and that is what makes the pattern reusable rather than lucky.
Read the outcome to decide whether to keep going. Read the diagnostics to decide what to change. Doing it the other way round is how teams end up with a library that performs beautifully in a report and flat in the account.