A founder asked us last month how many ads his agency should be producing for his Meta account. He was spending £85k a month. The agency was shipping six ads a fortnight. He thought that sounded fine because his previous agency had shipped four. He was wrong by roughly an order of magnitude.

This is the question nobody wants to give a straight answer to, because the straight answer is uncomfortable. So here it is. In a post-Andromeda world, the volume of creative that wins on Meta is not the volume of creative most brands are producing, and the gap is widening every quarter.

Why the “three to five ads per ad set” rule is finally dead

For years, the received wisdom was that you wanted three to five ads in each ad set so Meta could “learn” which one worked. The system would push budget toward the winner. Everyone was happy.

That logic depended on the old auction. The system was choosing between your three ads inside your ad set, against a shortlist of competitor ads from the broader pool. Three was enough to give Meta a signal without confusing it.

Andromeda doesn’t work that way. It evaluates your whole account-level creative pool against the entire eligible market for every impression. The “ad set” is a budget pool with a goal; it’s not a creative competition arena. The model isn’t picking between your three ads. It’s picking which of all your ads (and your competitors’) deserves this specific impression. Three is no longer enough surface area for the model to find the wins. Twenty starts to be.

The real maths of creative volume under Andromeda

Meta’s own published guidance now talks about “creative diversity” as a delivery factor. Translated out of platform-speak: the more genuinely different ads you have in the pool, the more chances the system has to match an ad to a user it would otherwise have shown a competitor.

A 50-word version for anyone in a hurry: at £20k a month spend, you need roughly eight to twelve fresh creative concepts a month to keep the system fed. At £50k, plan on fifteen to twenty. At £100k+ you’re into the twenty-five to forty range. These aren’t variants of one idea. They’re concept-level different. Variants on top of that.

The numbers come from what we see hitting fatigue inflection points across our accounts. A creative that started at a 1.4% CTR and a £22 CPA will, on most accounts above £50k a month, hit fatigue inside seven to ten days now. Two years ago it would have lasted three to four weeks. The system is metabolising creative faster because it’s metabolising everything faster.

What actually counts as “a different ad”

Here’s where most brands get fleeced. Their agency ships them ten ads a month and they think they’re well-fed. But six of those ten are the same creative with a different hook, the same hook with a different background colour, or the same video with a different end-card. The model doesn’t see ten ads. It sees one or two concepts in heavy rotation.

A “different ad”, in the way the model actually treats them, means a different opening three seconds, a different visual format, or a different angle on the offer. Same offer, four hooks: that’s one concept with four variants. Useful, but not what creative diversity is asking for.

If you want to know whether your creative pool is genuinely varied, look at the first three seconds of every active ad. If you could swap them around without anyone noticing, you have a volume problem dressed up as a quantity success.

The honest minimum by spend level

I’ll give numbers, because the style guide here is supposed to be honest about specifics. These are starting points, not absolutes, but they’re closer to what works than what most agencies will quote.

At £20k–£40k a month: eight to twelve new concept-level ads per month, with two to three variants of each. Roughly twenty-five live ads in the account at any time.

At £50k–£100k a month: fifteen to twenty new concepts per month, three to five variants of each. Forty to sixty live ads. This is also the spend level where the gap between agencies that produce in-house and agencies that outsource creative starts to become visible in the ROAS line.

At £100k+ a month: twenty-five to forty new concepts per month, with a properly resourced internal or embedded creative function. You’re effectively running a content studio. Brands that don’t accept that fact tend to plateau around £120k–£150k and then quietly slip back.

These numbers will sound mad to anyone who built their model around the old volumes. They are not mad. They are the cost of competing in an auction where the system is screening millions of ads for every impression. Showing up with six ads a fortnight is showing up to a knife fight with a butter knife.

What this changes about how you commission creative

The implication isn’t “find an agency that produces more ads.” It’s “redesign the brief so producing more is structurally possible.” That means modular shoots that yield ten cuts instead of one hero. UGC pipelines that ship weekly, not quarterly. Creators on retainer rather than per-project. AI tools (real ones, not the lazy generative-video sort) used to multiply variants once a winner is found.

If your agency tells you twenty creatives a month is overkill, ask them what they think the auction looks like underneath. The answer will tell you whether they’re a 2022 agency wearing a 2026 logo.

The brands that win Meta in 2026 will not be the ones with the cleverest single ad. They’ll be the ones with the most relentless creative engine. If yours isn’t built for that, the engine (not the ads) is what to fix first.

If your account is producing fewer ads than this and you’re wondering whether that’s why scale has stalled, We’re easy to reach.