How long brands keep an ad running on Facebook and Instagram

People who watch their competitors’ ads say the same thing in different words: the ads come and go fast, and they wonder whether their own are tiring early. Meta’s Ad Library shows when an active ad started running,⁵ but not what it spent or when it stopped, so nobody outside the account can count this. These three studies counted it from inside accounts.

All three come from companies that sell to advertisers, and each counted the accounts it can see: Interconnections its managed accounts, Motion its customers, Biddyco the accounts it has visibility into. Each publishes its method, so you can judge how far its numbers reach your own account.

What each study counted

They measure different things, so the numbers do not add up to one answer. The first two count every ad in their sample. The third starts from the ads that had already earned the most money, then watches what happens to them.

Most ads stop before they can tire

The usual question assumes an ad runs until the audience is bored of it. Interconnections found that assumption false for most ads: 80% never reached 100,000 impressions, and the top 10% of creatives carried 68% of the spend.¹ It says plainly that its sample leans towards ads that earned budget, so the true share that never got far is higher.¹

Motion’s larger count points the same way. It treats any ad switched off before 28 days as having lost, and about half of all creatives landed there.² Its window runs through the holiday promotions, and it warns that a steadier period may look different.²

So when a competitor’s ad vanishes after two weeks, the likeliest reading is not that it tired. It may never have earned much budget at all, and the ad library cannot tell you either way. It shows you what started and what kept going, and how to research competitor ads covers reading that without mistaking a long run for proof.

The ads that earn budget fade over months

Biddyco tracked each account’s top 5 creatives from June. A month later they kept a median 76% of their spend; by September, 29%. Of the top 10 creatives in those accounts in September, 46 of 70 had not been running at all in June.³

It also shows how much accounts differ: in one, 2% of September’s spend went to creatives from June, and in another, 88%.³ Seven accounts is a small sample, which makes this a shape to look for, not a number to plan by.

What it means for your own ads

None of these figures is a target. An ad does not tire because it reaches a certain age. Meta’s guidance is to watch frequency beside your results, and to consider new creative when results drop as frequency rises.⁴ Interconnections could not test the popular frequency thresholds at all, because its data had no reach at ad level.¹ How to tell when an ad has tired covers the signs, and a new opening, a new format or a new idea covers what to make next.

If most ads stop early and the few that last still fade, the work never ends: how many new ads brands launch a week shows what that looks like at each level of spend. When you find an ad a brand kept running, make your version of it from your own product and words.

What these numbers do not say

• Whether a long-running ad worked. These studies count spend and days; a brand can keep paying for many reasons.
• Why an ad stopped: a new offer, a season or a budget ends a run the same way a tired ad does.
• Anything outside Meta, or about accounts unlike the ones each company can see.
• That your account should match. Each study’s sample is its own customers or the accounts it can see.