How many Meta ad creatives to produce per month, and how often to replace them
Meta publishes no number of ad creatives to produce per month. Here is what our last two launches planned, how to count creatives without fooling yourself, and what the studies say about how long they last.
The short answer
Meta publishes no number of ad creatives to produce per month. Our last two launches planned 21 and 24 creatives for the first month, built on seven and six distinct ideas. After that, volume depends on how many results your budget buys each week. Replace a creative when its cost per result rises: the 2 to 4 week benchmark is only a starting point.
How many creatives do you need for a first month of Meta ads?
A creative is an ad image or video. People who contact us to produce them put the question this way: “how many creatives per month for this budget?” They also ask for examples and a test before committing.
Meta gives no figure. Its page on creative differentiation asks advertisers to vary their creatives without setting a number. Jon Loomer, who specializes in Meta ads, advises going beyond two or three ads, while pointing out that creating 50 guarantees no results.
Take two launches we prepared. For a software product for small businesses: 21 creatives across seven ideas. For an app aimed at independent workers: 24 creatives across six ideas, detailed in the table. These are first-month plans. They guarantee no results and do not transfer as they are to a different budget.
| Medium | Number planned | Why this number |
|---|---|---|
| Static images | 10 | Quick to produce, so several messages can be tested at once |
| Carousels | 5 | Several images to explain an idea or a use |
| Creator videos (UGC) | 5 | A person talking to camera, slower to obtain |
| Animated videos | 3 | Show the product without a shoot |
| Film | 1 | The longest to produce, so only one at the start |
What counts as a creative: an idea, a variant or a format?
This is what distorts comparisons between quotes the most. For a software client, we adapted one hook for 40 industries and three image formats: 120 files delivered, from a single starting idea. Presented as “120 creatives”, that batch would look far more varied than it is.
Meta explains that overly similar ads can be shown to the same groups of people, which limits their reach. Its creative fatigue page recommends a new image or video that is “materially different” from the original. Changing a button color does not meet that bar.
So count on three levels, and ask your provider how many they deliver at each one:
| Level | Example | What it brings |
|---|---|---|
| Idea | Win clients without hiring, or show the product in 30 seconds | A new message for new people |
| Variant | The same idea with another opening, headline or person | Separating several versions of an idea |
| Format | The same creative in 4:5, 1:1 and 9:16 | A proper display per placement, nothing more |
How do you match the number of creatives to your ad budget?
Start with a division: weekly budget divided by your cost per result, meaning the cost of a purchase or a sign-up. That gives you the number of results bought per week. This figure matters more than the budget alone, because Meta says an ad set exits its learning phase after about 50 results in the week following its last significant edit, a threshold also reported by specialists.
Meta then seems to contradict itself. Its learning phase page advises against high ad volumes, because the system learns less about each ad. Its April 2025 article instead presents creative variety as the most effective lever for finding the right people, and quotes a brand that went from 3 or 4 to almost 50 new creatives per week. We read the two together: variety is measured in different ideas, and a small budget cannot separate many of them at once.
The table gives our starting benchmark. It comes neither from Meta nor from a study: it uses the ratio from our launches, about three creatives per idea, and should be corrected with your own numbers. Under 50 results per week, we would also cut the first month to three ideas, so nine creatives. Campaign setup is a separate job and not the subject of this guide: decide this volume with whoever runs your campaigns.
| Results bought per week | New ideas per month | Creatives to produce, at three per idea |
|---|---|---|
| Under 50 | 1 or 2 | 3 to 6 |
| 50 to 200 | 2 to 4 | 6 to 12 |
| Over 200 | 4 or more | 12 or more |
How often should you replace your Meta creatives?
The most repeated figure is every 2 to 4 weeks. Segwise presents it as a starting point to confirm with your own data, and Superads points out that no pace works as a rule. Two recent measurements shed more light on the question.
- Interconnections, August 2026. Across 368 ads from 15 brands, the median creative stays live for 18 days, and 80% never reach 100,000 impressions. The author concludes that most stop because they do not work, before they have time to wear out. Stated limits: 15 brands, four months, and ad frequency not measured.
- Skaler, April 2026. Across 2.9 million inactive ads, the median duration is 3 days, and 17 days for those above 500,000 EU impressions. Many launched ads are simple tests, and the data collection method is thinly described.
Should you produce creatives continuously or in batches?
We deliver in batches, for a practical reason: each batch goes out with its ideas, and we know what it was testing when we read the numbers. Meta also says that creating an ad or significantly editing an ad set restarts the learning phase, and advises against unnecessary edits.
Before launch, we set the stop threshold with the client. Depending on the account, it sits at twice the target cost per result, or three to four times. A threshold decided in advance avoids cutting a creative on a hunch. The Interconnections study points the same way: the ads with the best click-through rates over the first three days ended up with fewer impressions than the lowest-ranked ones.
One shoot can be enough to produce several variants of a video. On one account, a creator filmed five three-second openings, with a different outfit, light and time of day, which we edited in front of the same screen recordings. This order replaces several shoots when only the opening of the video changes.
A new creative needs time before its results settle. If you are aiming for a sales peak, plan it ahead, like the rest of your Black Friday preparation.
How do you know a Meta creative is worn out?
Meta shows two statuses in the Delivery column. According to its help page, “Creative limited” signals a cost per result above your past ads but less than twice as high. “Creative fatigue” appears from twice that level, a definition also reported by Jon Loomer. This feature exists only for ad sets with a single creative: an account running several per ad set may never see these statuses.
A cost increase is not always fatigue. Jon Loomer describes accounts where the ad simply moves from past visitors to new people, without the creative being at fault. Before replacing anything, check these points:
- Cost per result. Interconnections reports +19.6% per acquisition between 500,000 and one million impressions, while click-through levels off after a 5.5% drop. In this sample, cost is a better alert than clicks.
- The audience reached. Check whether the share of past visitors has fallen. If it has, the creative may not be the cause.
- The landing page. More visits without more sales can come from the site: a real estate agency described this to us after two months of ads. These seven causes of an online store that does not convert help you check this point.
- What to do next. If the creative is truly worn out, Meta notes that keeping the original active may be better while adding a new ad, rather than pausing it.
What should you look at in competitors’ ads before setting your volume?
The Meta Ad Library is open to everyone and contains all active ads. It also shows the date an ad started running. An ad still live after 30 to 45 days, with several variants, is the most useful public signal we have found, since the library shows no results.
On October 26, 2026, we ran this check for a software publisher we work with, against eleven competitors. Our client had launched 23 ads in seven days, the others between 150 and 450 per week. The gap is striking, but these figures count ads, not ideas: on our client’s account, 372 ads came down to 219 distinct creatives.
The first lever we noted was simpler than producing more: finished creatives were still waiting to go live. Before raising your volume, check how many creatives already produced are sitting in a folder.
What should you ask a creative provider before signing?
These questions come from exchanges with companies that contacted us, sometimes because their previous provider no longer had time or because they were starting paid acquisition:
- How many distinct ideas are in the announced volume? The number of files means nothing without the number of ideas.
- Do you have examples in my format? A creator video and a static image do not take the same work.
- Can we run a test before committing? A first month works as a pilot: it shows the quality, the turnaround and the way of working.
- How long do the usage rights last? An agency told us it discovered ad rights limited to two months, with a high buyout to extend them. Have the duration and the uses written down before ordering.
- Do you use AI, and for what? One prospect told us they preferred a transparent method, and another looks first at voice and lip movement.
Where do you start when setting your creative volume?
Calculate your results per week, count the ideas you have already tested live, then list the finished creatives not yet running. These three figures are enough to set a first volume. If your ads send people to a store that also needs work, the topic also covers the store itself. To talk it through with us, write to us describing your situation.
Frequently asked questions
- How many Meta creatives should you produce on a small budget?
- A small ad budget calls for few ideas at a time, because it buys few results. First divide your weekly budget by your cost per result. Under about 50 results per week, the level Meta gives for exiting the learning phase, we would start with three ideas and nine creatives at launch, then one or two new ideas per month. This benchmark comes from our practice and should be corrected with your numbers. Producing 30 creatives on that budget would scatter the results: none would get enough for you to judge it. Keep the differences between ideas clearly visible, so that a result teaches you something: changing the message tells you more than changing a color.
- Should you stop a working creative because it is old?
- No: as long as its cost per result stays on target, an old creative can keep running. Age alone proves nothing. Jon Loomer writes that you do not need to refresh your ads every few weeks, provided you run them under the right conditions, and Interconnections recommends deciding on cost per acquisition rather than a fixed schedule. Meta points the same way: its creative fatigue page notes that keeping the original active while adding a new ad may be better. So produce what comes next before you need it, and add it without cutting what still pays.
- Should you start with videos or static images?
- No public Meta rule settles this, and the right mix depends on your product. In the first month of the app whose plan appears above, 15 of the 24 creatives were static images and carousels, and 9 were videos. We often start this way because a static image is quick to produce and lets you compare several messages. A creator video takes more time: you have to find the person, write the brief, wait for the shoot, then edit. Once a message stands out, video develops it. If your product only makes sense when seen working, reverse the order and start with a short demonstration.
- Can you test a creative provider before committing to a monthly volume?
- Yes, and this request keeps coming up among those who contact us about creatives. A useful test covers a full first batch, with a written scope: the number of ideas, the number of variants, the formats, the number of revisions, the duration of the rights. A single trial image teaches you little, because it shows neither whether the pace holds nor whether several ideas can be proposed. Also decide in advance what you will look at when the test ends, such as cost per result per idea, and over what period. A creative’s numbers over its first three days are not enough to judge it: the Interconnections study shows that this early ranking predicts the rest poorly.
- Do AI-generated creatives count toward the volume?
- Yes, if they carry a distinct idea. For Meta, the production tool does not make two ads different: its pages talk about overly similar creatives and ask for a new, materially different creative. AI makes it possible to produce more variants in less time, but a hundred versions of the same message remain one idea. The head of an online academy told us that voice and lip sync remained the weak point of AI-generated creator videos. Listen to videos with the sound on before running them, and ask your provider what is generated and what is filmed.
