Why do my Facebook ads sell for 3 days and then stop?
Usually because Meta finds the easiest buyers first: in the first days it shows the ad to the people most ready to buy, and once they have bought or scrolled past, it keeps paying to reach the same people again. Before you change anything, check that sales really stopped in your store and not just in Meta's report, look at frequency and cost per result together, rule out a change you made yourself, and count whether three good days were ever a trend.
The pattern is familiar to almost everyone who runs Meta ads for a small store. A new ad goes live, two or three orders arrive in the first days, and then nothing — same ad, same budget, same audience. It is tempting to conclude that the ad "broke" or that Meta "throttled" the account. Usually neither happened. Something ran out, or something was changed, or the first days were luck. The steps below tell those apart, so you fix the right thing.
Step 1: Check that sales stopped in your store, not just in Meta
Meta's report and your Shopify orders are two different counts. Meta credits a sale to the ad only if it can connect the order to someone who clicked or viewed it, and some of those connections arrive days late or never arrive at all.
- Count Shopify orders, not Meta purchases. Compare total orders in the three "good" days with the three days after. If store orders did not fall, the ads may still be working — only the credit Meta gives itself changed. Why Meta and Shopify report different numbers explains how that gap opens.
- Wait a day before reading yesterday. Meta often attributes a purchase to the day of the click, so yesterday's number can still rise. A "zero" from the last 24 hours is not yet a zero.
- Place a test order if purchases disappeared from Meta completely. A pixel or app that stopped reporting looks exactly like ads that stopped selling.
Step 2: Read frequency and cost per result together
Frequency is how many times, on average, each person in the audience has seen your ad. Cost per result is what each purchase costs you. Either number alone tells you little; the two together tell you whether the audience is used up.
- Frequency up, cost per result flat: probably nothing is wrong. Retargeting runs at high frequency on purpose, and a budget increase raises frequency too.
- Frequency up and cost per result up at the same time: the audience is tiring. The people most likely to buy have been reached, and the ad keeps going back to the same eyes at a rising price.
- Compare with your own account, not with a rule of thumb. "Frequency above 3 is bad" ignores how your account normally runs. What matters is whether both numbers have clearly left their usual range for you.
Step 3: Check whether you changed something on day 3
Many "it stopped" stories start with a change made because it was working. Meta treats a significant edit as a new start, and a new start means relearning.
- A large budget jump. Doubling the budget overnight because the first days sold well asks Meta to find many more buyers at once — and they are, by definition, less ready than the first ones.
- Edited ads, audiences or the optimisation goal. Changing the text, the creative, the targeting or the conversion event sends the ad set back into learning, where results are usually unstable.
- Duplicating the ad set to "scale" it splits the same audience between two ad sets that now compete with each other.
If you made one of these changes, the drop may be the cost of the change, not a sign that the ad is finished. Give it a few days before you change anything else — one change at a time, so you can tell what helped.
Step 4: Check how big the audience really is
The smaller the audience, the sooner it runs out. A narrow interest list, a small country or a lookalike built from very few customers can be exhausted within days at a modest budget.
- Look at the estimated audience size in the ad set. If it is small and the frequency climbs quickly, the audience is the limit, not the ad.
- Try broader targeting with the same creative. Meta usually finds buyers better when it is given room to look.
Step 5: Test a new creative before a new audience
When the audience is tiring, there are two options: show the same people something genuinely different, or show the same ad to new people. The first test is usually cheaper and tells you more.
- Make it genuinely different — a new hook in the first seconds, a different angle, a different format. A new colour or a new headline on the same video is seen as the same ad.
- Keep the audience unchanged for this test, so the result tells you whether the creative or the audience was exhausted.
- Plan for it. An ad that sells for a while and then fades is normal, not a failure. Stores that run Meta steadily keep the next creative ready before the current one wears out.
Step 6: Count before you call it a trend
Three sales in three days, then none in three days, can simply be chance. With only a handful of orders, a few good days and a few bad days look like a story that isn't there.
Judge in money, not in days: let the ad spend two to three times what one customer is allowed to cost you, then compare its cost per sale with that limit. Is $20 a day enough for Meta ads, and when should I stop? works the arithmetic through step by step.
KPIHelm's SUYLA plan checks this every morning for Meta and TikTok, against your own account's normal rather than an industry figure. When the same people are being shown your ads clearly more often than usual and each result is costing clearly more than usual — both at once — it tells you the audience is tiring and suggests testing a genuinely different creative before widening the audience. One of the two alone does not trigger it, and it needs at least ten days of your own history before it judges. Install it free on the Shopify App Store or see a sample report.