A view is not a result. What decides whether a budget is renewed is the ratio between what a creative cost and what it brought in while it was alive.
The three figures that count
They are calculated in this order, and none replaces another.
Cost per usable creative: what you spent in total, divided by the number of videos actually fit to run. Videos delivered but unusable count in the cost, not in the result, and that is what makes this figure explode for brands that brief badly.
The acquisition cost obtained with that creative, compared with your usual average. It is the only figure that says whether the video does better than what you already had.
And the lifespan, described above.
What to measure at each stage
Three indicators, each for a different question.
Three-second retention judges the hook, nothing else. A poor figure here means the opening has to change, not the video.
The click rate judges the promise. If it is good while retention is weak, your hook is attracting the wrong people.
And the purchase rate on the page judges the consistency between the video and the landing page. If it drops, the problem is almost never the video.
That separation avoids the most frequent mistake: throwing away a good video because of a bad page.
How to build an honest test
Four rules, and they are demanding.
One variable at a time: same product, same page, same audience, same budget. Otherwise you are measuring nothing.
A budget large enough to escape chance. A creative judged on a few dozen impressions teaches you nothing.
A duration fixed in advance, the same for every creative in the test.
And a decision written down before starting: at what figure you stop, at what figure you continue. Without that, the result will be read in the direction of what you were hoping for.
The most frequent measurement mistakes
Four, and they cost more than bad videos.
Comparing a creative launched in the middle of a sale period with one launched in an ordinary month.
Killing a creative before it has reached the minimum budget decided at the outset.
Attributing to the video a result obtained by a discount announced at the same time.
And changing the landing page during the test, which invalidates everything.
What you record, and when
| Indicator | When to read it | What it decides |
|---|---|---|
| Share still there at 3 s | The next day | Keep or change the hook |
| Clicks per thousand views | At three days | Keep or change the ending |
| Cost per order | At two weeks | Continue or stop |
What is written before publishing
The goal, the format, the hook used and the landing page: without those four lines, a figure cannot be read.
What is never compared
Two videos posted on different dates, to different audiences or with different budgets.
What is specific to Brazil
Two points shift the benchmarks.
The commercial calendar is highly concentrated: a few dates absorb a significant share of annual demand, and a creative tested during those peaks compares with nothing.
And instalment payment changes the buying decision: the question becomes commitment over time, not price. So also measure the delay between first view and purchase, which is often longer here than the dashboard assumes.
On price and financing statements, the wording used comes from the brand's approved documentation, and the applicable obligations are checked with the competent authority and with a lawyer before distribution.
