How Much Should a Facebook Ad Spend Before I Judge It?

You launched five ads on Tuesday. By Thursday one has spent $60 with no purchases and another has two sales, and someone wants to know which to cut. At that volume the gap between them is mostly luck, and cutting on luck kills ads that would have worked.

Judge a Facebook ad by what it has spent relative to your target cost per result, not by days or a fixed dollar amount. Let it spend two to three times your target CPA before you cut it for not converting. Before you call it the winner, it needs far more: enough for each ad to collect a few dozen conversions.

Why spend beats days as the yardstick

Day-count rules like "give it three days" or "wait a week" are popular because they are easy to follow. But a day is not a unit of evidence. An ad on $10 a day and an ad on $200 a day have learned completely different amounts by day three.

What you are actually waiting for is conversions, and you buy conversions with spend at roughly your cost per result. So the honest unit of patience is your target CPA. The question to ask is how many times over this ad has spent what one result is supposed to cost.

That multiple also travels across accounts. A $2,000 product with a $400 target CPA and a $30 subscription box with a $25 target CPA cannot share a dollar rule, but they can share a multiple.

Time still matters at the edges. Give any test a few full days so weekday and weekend behaviour both show up. The first stretch of delivery is the platform exploring who to show the ad to, not yet settling on an audience. Once both conditions are met, spend is what tells you whether you have enough evidence to decide.

How much spend proves an ad is losing?

The kill decision is the cheaper one, because it asks a narrow question: is this ad plausibly hitting target at all? Basic probability answers it.

Suppose an ad really were converting at exactly your target CPA. Spending one times target should then produce one conversion on average. If you treat conversions as random, independent events, the odds of getting zero by chance are about 37%. At two times target CPA, the chance of zero drops to about 14%. At three times, it is about 5%.

That is why our working rule is two to three times target CPA. At two times, a zero-conversion ad is probably not hitting target. At three times you can be fairly confident it is not. Cut earlier and you will regularly kill ads that were fine, just unlucky in their first few hundred impressions. Hold on much longer and you are paying to confirm what you already know.

The same logic covers ads that convert, but badly. One conversion at three times target CPA means the ad is running at roughly triple your target cost. That is not proof it can never work, but it is enough to stop funding it while better ads wait for budget. Zero-conversion ads raise their own checks before you pause, such as tracking, landing page and audience size. Those are covered in when to turn off a Facebook ad with no conversions.

How much spend proves an ad is winning?

Picking a winner asks a harder question: is ad A genuinely better than ad B? Small counts cannot answer it. Four conversions against two looks like double the performance, but at those numbers the ranking flips routinely on a rerun. The difference you are trying to detect is smaller than the noise around it.

To separate two ads with any confidence, each typically needs a few dozen conversions, and more if the true gap between them is modest. Translate that into money with your CPA. At a hypothetical $40 target, 30 conversions per ad means roughly $1,200 per ad before the comparison tells you much. That is far more than most testers allow, which is why so many early "winners" fall apart once scaled.

There are two practical responses:

• Look for a large gap, not a perfect one. Say one ad is running at half the CPA of the others, at meaningful volume. You do not need textbook statistics to back the obvious leader.
• Use earlier signals when conversions are thin. Click-through rate, cost per landing page view and add-to-cart rate arrive faster than purchases. They only correlate with revenue imperfectly, though. They tell you which creative earns attention, not which one earns money.

Whether a proven ad should then get more budget is a separate decision. It belongs to the wider question of which Facebook ads to turn off, keep or scale.

When should you move the threshold?

The multiple holds in most accounts, but four situations change how you apply it.

Long consideration cycles. For high-ticket products, a conversion can land days or weeks after the click. Hit your spend threshold, then wait for conversions to catch up before reading the result. Alternatively, judge on an earlier event that reliably predicts the sale.

Edits mid-test. Substantial changes send the ad set back into learning: new creative, new targeting, a different optimisation event. Spend before the edit tells you about the old version, not the new one. Decide what you are testing, then leave it alone until it has spent enough to judge.

Several ads sharing one ad set. Meta concentrates delivery on whatever it predicts will perform, so some ads barely spend. An ad that received $8 has not been judged; it has been skipped. You can accept the platform's preference as a signal. If you need a fair head-to-head, test in separate ad sets with matched budgets.

Small, warm audiences. Retargeting pools convert cheaply and hit the threshold fast, but performance often decays as the same people see the ad repeatedly. That decline is fatigue, a different problem from whether the ad worked in the first place.

How to set your Facebook ad testing budget

Work backwards from the thresholds. The number of ads multiplied by your kill threshold gives you the minimum spend just to clear losers fairly. Take five ads at a $40 target CPA with a three-times threshold. That is $600 to get an honest keep-or-cut read on all five. A winner verdict for the survivors pushes the figure much higher.

If that number is more than you can spend, test fewer ads, not cheaper ones. Two ads with enough budget to judge beat six with too little. You learn something definite instead of six things you cannot trust. Moving the optimisation event up the funnel is the other lever, at the cost of judging on a proxy.

Write the thresholds down before launch: the spend at which each ad gets cut, and the conversion count at which you will compare survivors. The exact rule matters less than whether you stick to it on Thursday morning, when an early number is tempting you to act.

Conclusion

The answer depends on which decision you are making. Cutting an ad is cheap to get right: two to three times target CPA with no conversions is enough evidence. Crowning a winner is expensive: each ad needs a few dozen conversions, or a gap so large it survives the noise. Days and fixed dollar rules feel safer because they are simple, but they measure the calendar rather than the evidence. Set the multiples before launch, budget for them honestly, and hold to them when the early numbers tempt you.

See how Execute adjusts live budgets continuously within the spend and scope limits you set, rather than waiting for a weekly review to act on them.