The single most useful rule for a Meta ads account structure is not a number of campaigns or ad sets, it is a volume of signal: each ad set needs roughly 50 conversions a week to leave the learning phase. Every structural decision, how many ad sets to run, whether to consolidate, when to split budget, follows from that one constraint. Fragmenting a fixed budget across many ad sets pushes each below the threshold, so none of them ever stabilises.

That is why modern account structure trends toward consolidation, not toward ever finer segmentation. Meta’s own delivery has moved the same way. Its next-generation ad retrieval engine, Andromeda, launched in 2024 with a reported 6% recall improvement and 8% ads quality gain on selected segments, and it rewards broad audiences and creative variety over the manual micro-targeting that used to justify a dozen ad sets.

This article explains the structure that follows from the way Meta now delivers ads: fewer ad sets, more signal per ad set, and consolidation as the default rather than the exception. It is a foundational lever for any B2B paid acquisition programme.

What does a Meta ads account structure actually look like?

A Meta account has three levels, and each one answers a different question. The campaign sets the objective, for example Sales or Leads. The ad set decides who sees the ads, where, and on what budget. The ad is the creative itself, the thing a person actually sees in their feed.

Account structure is the set of choices you make across those three levels: how many campaigns, how many ad sets under each, how the budget is split, and how audiences are divided. There is no single correct template, because the right structure depends on one variable that changes from account to account: your weekly volume of conversions.

That is the point most structure guides miss. They present a fixed diagram, three ad sets here, five ads there, as if it were universal. In practice, the same diagram that works for an account with 800 conversions a week starves an account with 60. Structure is arithmetic before it is a template.

Why does consolidation beat fragmentation?

Because the algorithm learns per ad set, not per account. The learning phase is the period after a significant change when Meta’s delivery system is still calibrating, and delivery stays volatile until the ad set gathers enough conversions to stabilise. Meta’s help centre documents a threshold of roughly 50 conversions in the week after an ad set’s last significant edit. Those official pages do not render reliably on direct access, but the figure is corroborated by multiple agency guides, including Modern Marketing Institute and AdControlCenter.

Now look at what fragmentation does to that number. If your account produces 60 conversions a week and you split it across five ad sets, each ad set averages 12, well under the threshold. Every one of them stays in learning limited, delivery never settles, and your cost per result stays high. Pool the same 60 conversions into one or two ad sets, and at least one clears the threshold and stabilises.

Fragmented versus consolidated ad set structure on MetaBar chart comparing two Meta account structures that spend the same weekly budget. On the left, a fragmented structure with five thin bars, each representing an ad set receiving about 12 conversions per week, all of them well below a horizontal dashed threshold line set at 50 conversions per week. Each of these five ad sets is labelled learning limited because it never reaches the volume the algorithm needs. On the right, a consolidated structure with two tall bars, each representing an ad set receiving about 58 conversions per week, both rising above the same 50-conversion threshold line and therefore able to leave the learning phase and stabilise. The chart shows that pooling the same budget into fewer ad sets is what lets each ad set clear the volume threshold.Same weekly budget, two structuresWeekly conversions per ad set. The algorithm learns per ad set, not per account.Threshold: about 50 conversions per week to leave the learning phaseFragmented: 5 ad setsabout 12 each, all learning limitedConsolidated: 2 ad setsabout 58 each, both stabiliseabove thresholdabove threshold
Illustration of the core trade-off. The same weekly budget, split five ways, leaves every ad set in learning limited; pooled into two, both clear the roughly 50-conversion threshold and stabilise. Source : Learning-phase threshold documented by Meta and corroborated by agency guides

The effect shows up in real accounts. One apparel brand cited by The Interconnections consolidated from 16 ad sets across five campaigns down to two Advantage+ campaigns in mid-2025 and cut its cost per acquisition by 23% over the following six weeks, with no change to creative. The gain came from pooled conversion signal, not from a better ad. That is the mechanism consolidation exploits, and it is why the modern delivery system, Andromeda included, favours simpler structures where the algorithm rather than manual segmentation decides who sees each ad.

How many ad sets per campaign should you run?

The right number of ad sets is the largest number that keeps every ad set above roughly 50 conversions a week. That formulation replaces any fixed figure because it builds in the only variable that matters, your conversion volume.

Work it in three questions, in this order:

  • What is your total weekly conversion volume? This is the numerator for every structural decision.
  • How many ad sets can that volume feed above 50 conversions each? This is your real ceiling, and for most B2B accounts it is one, two, or three.
  • Does each additional ad set add a distinction worth the cost? A genuinely different audience, geography, or funnel stage. If not, the split spends signal without buying anything.

For most accounts the practical answer lands between one and three ad sets per campaign. The platform’s own ceiling runs to dozens of ad sets, as Vizup notes, but that is a technical limit, not a recommendation, and almost no account should approach it. Only move to two or three when you can name the reason: prospecting versus retargeting, distinct countries that need separate spend, or a controlled test you actually intend to read.

The same arithmetic governs how many creatives you load into each ad set, and it interacts with how fast those creatives wear out, which is why watching creative fatigue metrics matters as much as counting ad sets.

When does audience overlap start cannibalising your budget?

Audience overlap is the second structural trap, and it is invisible on a campaign diagram. When two ad sets target overlapping people, Meta enters both into the same auction for the same impression, so you end up bidding against yourself and inflating your own CPMs.

The working thresholds, drawn from agency benchmarks such as AdsX and Bestever, are practical rather than official. Under 15% overlap between ad sets is generally fine. Between 15 and 25% is worth monitoring. Above 25%, overlap materially cannibalises delivery, and AdsX estimates the extra competition can inflate CPM by 18 to 25% on the overlapping inventory alone. Meta’s own audience overlap tool lets you check this before you launch.

There is a nuance worth keeping. High overlap between two broad prospecting audiences of similar size, say a 1% and a 2% lookalike, is normal and largely harmless, because both are doing the same job. The costly overlap sits at the boundary between prospecting and retargeting, where two ad sets chase the same warm user with different intent. The fix is either exclusions or, more often, consolidation of the overlapping ad sets into one.

CBO vs ABO: when to use each

CBO and ABO describe where the budget lives. With ABO, ad set budget optimisation, you fix the budget on each ad set, so every audience or creative gets an equal, protected share. With CBO, now called Advantage+ campaign budget, you set one budget at campaign level and let Meta move money in real time toward whatever converts most cheaply.

The naming has shifted under many advertisers’ feet. As Segwise documents, Meta merged its manual and Advantage+ campaign flows and made AI-driven budget optimisation the default for new campaigns. So the question is no longer whether to use CBO, it is when to override the default. Many accounts still debate switching to CBO when their campaigns are already on it.

The usable rule that survives the renaming is a division of labour:

  • Use ABO to test. When you want clean, comparable data on a new creative, hook, or audience, fixing the budget per ad set stops Meta from starving a variant before you can read it.
  • Use CBO to scale. Once you know which creative and audience win, one campaign budget lets Meta concentrate spend on the winners without you micromanaging each ad set.

CBO also pairs naturally with consolidation, because a single campaign budget across two or three ad sets is exactly the pooled structure the learning phase rewards. If you want the detail on how Advantage+ automation changes what you control, see our note on Meta Advantage+.

Should I consolidate my campaigns? A short decision test

Consolidation is not automatic, so run these four checks before you merge anything:

  • Are any ad sets below about 50 conversions a week? If yes, they are in learning limited and are candidates to merge upward.
  • Do any two ad sets overlap above 25%? If yes, they are bidding against each other and should be combined or excluded.
  • Does each ad set represent a real, budgeted distinction? If two ad sets exist only because someone split them out of habit, they belong together.
  • Are you testing, or scaling? Testing can justify temporary ABO splits you plan to read and close. Standing structure should trend toward consolidation.

The common failure is treating structure as permanent. A test structure that made sense at launch quietly becomes a fragmented standing structure that starves every ad set. Reviewing the account against these four checks, not redrawing a template, is what keeps the structure aligned with the volume it actually has. Structure also interacts with how budget is divided across platforms in the first place, which we cover in our guide to splitting budget between Google and Meta for B2B, and with what happens to efficiency as you push spend up, covered in why ROAS drops when scaling.

In short

  • Structure follows signal, not a template. Each ad set needs roughly 50 conversions a week to leave the learning phase, so the right number of ad sets is the most you can keep above that line, usually one to three for a B2B account.
  • Consolidation is the modern default. Pooling budget into fewer ad sets clears the volume threshold and fits how Meta now delivers ads through its Andromeda retrieval engine. One brand cut cost per acquisition by 23% on structure alone.
  • Watch overlap and know your budget mode. Above 25% audience overlap you bid against yourself; test with ABO for clean data, then scale winners with CBO, now the Advantage+ campaign budget default.

Account structure is a volume problem before it is a design problem, and the fix is rarely a prettier diagram. If you want to hold your current structure up against the volume it actually produces, book a diagnostic: we start from your weekly conversion count and your overlap, not from a template copied off a blog.