How often should you refresh Meta ad creative? (by format and account age)

There is no single "every X weeks" answer, and anyone who gives you one is selling a calendar, not a strategy. Refresh cadence is a function of three things: format, spend level, and account age. As a rough starting point, Reels tend to fatigue in about 7 to 14 days at meaningful spend, broad Feed prospecting often lasts about 3 to 6 weeks, and static tends to sit in the middle. But those are defaults you adjust, not rules you follow. The real trigger is the ad's own fatigue signals, not the date on the calendar.

Let me break down why each of the three levers moves the number, then give you a table you can actually use.

Format sets the base rate

Format is the biggest driver because it decides how fast the audience burns through the ad. Short vertical video is consumed quickly and repeatedly. A Reel in the feed gets watched, skipped, and re-served fast, so the same audience sees it many times in a short window. That compresses the fatigue timeline. Static and simple image ads are lighter to process, so a strong static concept can hold longer, especially in retargeting where the audience is small and warm but intent is high.

Placement matters inside format too. A Reels placement fatigues faster than the same video cut for the Feed, because the surrounding content and the viewing behaviour are different. So do not treat "video" as one bucket. A 6 second hook running in Stories and Reels is a different lifespan problem than a 30 second Feed explainer.

Spend compresses the timeline

Here is the part most cadence advice ignores: spend changes everything. Frequency is what actually kills a creative, and frequency is a function of how much you spend against how big the audience is. Pour more budget into the same audience and you hit the same people more times per week. The creative does not get worse. Your audience just gets tired of it faster.

So two accounts running the identical Reel can have completely different refresh needs. The one spending 500 a day into a 2 million person audience might be fine for three weeks. The one spending 5,000 a day into the same audience will see frequency climb and CPMs rise inside a week. When you scale spend, assume you have shortened the creative's life, and have the next batch ready before you push budget, not after.

Account age: favour stability in the first 90 days

New accounts and fresh launches are the exception to all of this. In roughly the first 90 days, the algorithm is still learning who your buyer is. Every time you swap creative aggressively, you fragment that learning and reset signals the system was starting to trust. Over-rotating early looks like diligence but it actually slows the account down.

So early on, favour stability. Let winners run longer than your instinct says. Resist the urge to refresh on schedule just because a calendar told you to. A mature account with months of conversion history absorbs new creative smoothly because the targeting is already well trained. A three week old account does not have that cushion. Same logic applies to a brand new campaign inside an old account: the campaign itself needs time to settle before you start churning creative through it.

One more rule that holds at every account age: launch new ads instead of editing live ones. Editing a running ad resets its learning. If you want a fresh variation, ship it as a new ad and let the old one wind down on its own.

The table: refresh ranges by format and placement

Use these as starting ranges at meaningful spend. Shorten them as spend rises, and lengthen them in the first 90 days.

| Format and placement | Typical refresh range | Notes |

| --- | --- | --- |

| Reels and Stories (prospecting) | 7 to 14 days | Fatigues fastest, watch frequency closely |

| Short video, Feed (prospecting) | 2 to 4 weeks | Slower than Reels, hook still decays first |

| Static and image (prospecting) | 3 to 5 weeks | Holds longer if the concept is strong |

| Broad Feed prospecting (any format) | 3 to 6 weeks | Large audience buys you time |

| Retargeting (any format) | 2 to 4 weeks | Small warm audience, frequency climbs fast |

| First 90 days, any format | Lengthen the above | Favour stability, do not over-rotate |

Watch the signals, not the calendar

Here is the mindset shift. A calendar cannot see your account. It does not know your frequency crept to 4, your CTR halved, your CPA drifted up while spend held flat, or that your CPM jumped because the same people keep seeing the same thing. Those are the numbers that tell you a creative is done, and they show up on their own timeline, not on a Tuesday every third week.

So build the habit of reading the fatigue signals to watch day over day, and refresh when the ad tells you it is tired. That is also why a real creative testing framework matters more than a schedule: you want 3 to 5 fresh variations on your winning concept ready every 2 to 3 weeks, so when a signal turns you have a proven replacement, not a scramble.

This is exactly the kind of day-wise reading that is hard to do by eye across a full account, which is where a tool like AdBrain helps: it reads the account data daily and flags what to scale, refresh, or kill, with a reason for each call, so the refresh decision is driven by the ad's own numbers instead of the calendar.

The short version: format sets your base cadence, spend compresses it, account age stretches it in the first 90 days, and the ad's own fatigue signals make the final call. Ranges get you in the ballpark. The signals get you the timing right.