When to Kill a Meta Ad in a Creative Test

Kill an ad when the spend is enough to judge it, not when you feel nervous. Below are the rules, with example numbers you can swap for your own, and the mistakes that cost the most.

The short answer

Kill an ad early only when something is broken or the spend has clearly proven it can't work. Otherwise wait for the verdict window. Most wasted creative budget comes from two opposite habits: switching off ads after one bad day, and letting a dud run for three weeks because nobody owns the decision.

The fix is the same for both. Write the rule before the test starts, in numbers, and apply it the same way to every ad. A rule you wrote while calm beats a judgment you make while staring at a red column.

This page gives you three layers: a same-day kill list, an early-kill rule tied to your target CPA, and a 7-day verdict for everything else.

Decide what you are killing: the ad or the concept

If you test one ad at a time, you learn about one execution. If you test one concept in one ad set, with a few variations of the same idea, you learn about the idea. That is the unit worth judging.

Group each test as a numbered batch: one concept, one ad set, a handful of ads. Then ask a better question than whether this ad failed. Ask whether this desire, angle and awareness stage failed, or just this hook.

One weak ad inside a batch whose siblings are doing fine is a small decision. A whole batch that misses is a bigger signal, and it should change what you brief next.

Same-day kills: no waiting required

Some ads should go off immediately because no amount of data will fix them. Check these in the first 24 hours:

• The ad is disapproved or limited, or it's running with the wrong page.
• The link goes to the wrong product, a broken page or a page that doesn't continue the ad's promise.
• Your purchase event isn't firing, so the numbers can't be trusted either way.
• The ad has spent a meaningful amount and earned almost no link clicks, which usually means the first frame or the audience match is off.
• The copy makes a claim you can't support. Compliance risk is not a metric to wait on.

Early-kill rules tied to your target CPA

Early kills need a spend floor. The floor should be a multiple of what a purchase is worth to you, not a calendar day. The numbers below are an example: a target CPA of $40, an average order of $80, break-even ROAS of 2.0. Replace them with yours.

Rule one: until an ad has spent 2x your target CPA ($80 in this example), do not judge it on ROAS or CPA at all. One purchase more or less swings the result too much.

Rule two: if an ad reaches 3x target CPA ($120) with no purchases and no add-to-carts, pause it. It has had a fair chance to show any buying intent.

Rule three: if an ad has passed 2x target CPA, its CPA is more than double the target, and its click-through rate is the lowest in the batch, pause it. Both the cost and the early signal point the same way.

Rule four: never kill on a single day. Look at the span since launch, not yesterday.

Low-volume brands should scale the floors down only with care. If a purchase takes a long time to arrive, base the floor on a higher-funnel event you've shown predicts sales, and say so in writing.

What not to kill on

These look like reasons to pull an ad. Mostly, they aren't.

• Day-one ROAS. Early numbers are noise until spend passes your floor.
• A high CPM on its own. If the ad converts at an acceptable cost, the CPM is just the price of the audience.
• A low click-through rate alone. A narrow ad can click less and still sell to the right people.
• Low spend. If Meta gave an ad little budget, that is not proof it's bad. If it beat the campaign's ROAS on that small spend, it deserves a second look.
• Comments you dislike. Judge ads on sales and cost, then deal with the comments separately.
• A fresh edit of a loser. If the concept missed, a new hook on the same idea is not a new test.

The 7-day verdict for everything else

Ads that survive the early rules get one decision after seven days, judged against the campaign they ran in, not against a number from another account. Hitrate uses one written rule for this, and you can copy it into a sheet.

Scaler: the batch took a big share of its campaign's spend, and the campaign spent at least 10% more per day than the week before. Spender: a big share of spend, but no growth. Efficient: it beat the campaign's own ROAS but got little spend. Miss: none of those.

A big share means 30% of the campaign when it spends under $1,000 a day, 20% under $5,000 a day, and 10% above that. Adjust the numbers to how your account behaves.

What to do with each label: a Miss gets switched off and logged. A Spender stays on and gets variations. An Efficient batch is not a failure, so test it again with more budget or in a different campaign. A Scaler gets the next brief built on it. A batch can still become a Scaler in week 2 or 3, so don't bury a Spender or an Efficient batch too early.

If you judge on cost per acquisition instead of ROAS, set a target CPA and use it as the baseline. The method doesn't change.

Hitrate runs this for you. It names each ad by batch so results map back to the right concept, syncs from Meta daily (read-only) and applies the labels after seven days. The labels come from the written rule, not from a model.

Log the reason, or you'll repeat the mistake

Every kill needs one line: what was the concept, why was it switched off, and which rule fired. Without it, the same angle comes back in three months under a new name.

Then look across batches, not ads. Hit rate is the share of tested batches that became a Scaler or a Spender. Split it by desire, angle, awareness stage and format. If problem-aware hooks hit twice as often as solution-aware ones in your account, the next brief should know that.

Hitrate tracks hit rate by those four cuts automatically, because the ad names carry them. In a sheet, it takes four extra columns and about ten minutes on a Friday. Skipping it is the real cost of weak kill rules.

Related guides

• The creative strategist workflow, from research to learnings
• How to identify creative fatigue in Meta ads (with thresholds)
• Meta hook rate benchmark: what counts as a good hook rate
• How to improve hook rate on Meta video ads
• A Meta ads creative testing framework you can run weekly

Updated 2026-10-02