TL;DR
Nobody ever lost money because their Meta ads account structure was too simple. In 2026, consolidated structures win: one prospecting campaign, one retargeting campaign, and a testing setup that doesn't starve your winners. Every split fragments learning and thins your signal, so each additional campaign needs to justify its existence. Most agencies build cathedrals with 40 ad sets because complexity looks like expertise on a slide deck. It isn't. Simple structures with great creative beat complex structures with average creative every single time.
Your account structure is not the strategy
Let's get this out of the way: account structure is plumbing. It matters, but it's maybe 10% of performance. Creative is 60%, offer and economics are 25%, and the remaining 5% is everything else including structure, bid strategy, and which day of the week you launch (it doesn't matter, stop asking).
And yet structure is where we see the most theatrical over-engineering. Audits come in with campaigns named things like "US_Prospecting_Broad_25-44_Male_ExcludeEngagers_v7_FINAL" and we have to deliver the bad news: this isn't sophistication, it's clutter. Every ad set you add is another mouth splitting the same budget, another learning phase that never finishes, another fragment of signal too small for Meta's system to do anything intelligent with.
The 2026 reality is simple. Meta's delivery system is better at finding buyers than your interest stack is. Its job gets easier with consolidated data and harder with fragmentation. So the correct structure is the simplest one that still respects genuinely different jobs to be done.
The structure we'd build from scratch
Three campaigns. That's the whole thing for most accounts.
Campaign 1: Prospecting. One CBO campaign, broad targeting, all your prospecting creative inside it. Budget gets the lion's share, usually 70-80% of total spend. The ad sets inside exist for genuinely different creative angles or formats, not for audience slices. Meta figures out who converts. Let it.
Campaign 2: Retargeting. One campaign, smaller budget, 10-20% of spend. Website visitors, video viewers, engagers, email list. Keep it boring. Retargeting is a reminder, not a second full-funnel strategy, and anyone running six retargeting campaigns is just paying to follow people around the internet.
Campaign 3: Testing. This is the one people argue about. Either a dedicated testing campaign with a fixed budget (10-15% of spend) where new concepts prove themselves before graduating to prospecting, or test inside the prospecting campaign with new ads added to existing ad sets. Both work. What doesn't work is having no testing home at all, which means new creative either never launches or gets thrown into the prospecting campaign where it gets crushed by established winners before it has a chance.
That's it. Three campaigns. If your account does under $50k a month, you could arguably run two.
When splitting actually makes sense
Simple doesn't mean one campaign for everything forever. Split when the split reflects a genuinely different job:
Different economics. Separate geos with wildly different CPAs and margins deserve separate campaigns, because one budget pool will always starve the expensive geo or flood the cheap one. Same for separate product lines with different price points and break-even math.
Different stakeholders. If the client needs a hard budget split between brand and performance, or between two regions with separate P&Ls, structure follows the money. That's a business requirement, not a media tactic.
Different objectives. Lead gen and purchase optimization are different jobs. Don't force them into one campaign and hope.
What doesn't justify a split: twelve interest-based ad sets ("but what if moms 25-34 convert differently than moms 35-44"), separate campaigns per creative ("Video Campaign" vs "Static Campaign"), or anything with the word "funnel" that assigns 3% of budget to a stage nobody can measure. If you can't articulate why the split exists in one sentence without saying "to test," it probably shouldn't exist.
What most agencies overcomplicate
The lookalike zoo: seven lookalike percentages of five different seed audiences, each in its own ad set, each spending $8 a day, each permanently in Learning Limited. In 2026 this is astrology. Broad prospecting with good creative will find those people and then some.
Micro-segmentation: splitting audiences so finely that no ad set ever exits learning. Meta needs roughly 50 conversion events per ad set per week to optimize properly. Do the math on your budget before you slice it into eight pieces. Most accounts we audit are running ad sets that mathematically cannot succeed, then blaming creative.
The "full funnel" with no budget: awareness, consideration, conversion campaigns where awareness gets $10 a day. That's not a funnel, it's a participation trophy. Either fund the upper funnel properly or admit you're running a conversion-only account, which is fine.
Naming conventions longer than the ad copy. If your campaign name needs a legend, the structure is too clever.
The budget minimums nobody respects
Here's the unsexy math that kills most complex structures. An ad set needs about 50 optimization events per week to exit the learning phase and stabilize. If your CPA is $40, that's $2,000 a week per ad set just to give Meta enough signal. Running ten ad sets on a $3,000 weekly budget means every single one is learning-limited forever, and you're essentially paying Meta to guess.
So before adding an ad set, ask: does this ad set have enough budget to get 50 conversions a week? If not, consolidate. This one rule would fix half the accounts we've ever audited.
Simplifying an existing mess
If you're staring at a 40-ad-set account right now, don't nuke it overnight. Migrate:
- Identify your 3-5 best-performing ad sets by actual conversion volume, not ROAS on tiny spend.
- Build the new consolidated prospecting campaign alongside the old one.
- Move winning creative into the new structure, keep budgets modest at first.
- Once the new campaign is stable and spending, pause the old fragments one by one.
- Resist the urge to "keep that one small ad set, it's doing okay." It's doing okay on $12 a day. Let it go.
Give it two weeks. The account will be calmer, cheaper to manage, and usually performing better, because you finally stopped splitting your signal into confetti.
FAQ
How many campaigns should a Meta ad account have in 2026? For most accounts: three. One prospecting, one retargeting, one testing. Add campaigns only for genuinely different economics, geos, or objectives.
Is CBO or ABO better for the prospecting campaign? CBO, in most cases. It shifts budget to what's working in real time instead of locking spend into ad sets you've pre-judged. ABO still has its place for forced testing.
Should I separate creative by campaign (video vs static)? No. Put all formats in the same prospecting campaign and let delivery sort it out. Separating by format is structure theater.
How much budget does each ad set need? Enough for roughly 50 optimization events per week. At a $40 CPA, that's about $2,000/week per ad set. Below that, consolidate.
When should I split prospecting by audience? Almost never in 2026. Split by audience only when the audiences have genuinely different economics or when business requirements demand separate budgets. Interest slicing for its own sake just fragments learning.
Does account structure really matter that much? Less than creative, offer, and economics. But a bad structure actively harms performance by fragmenting signal, while a good structure just gets out of the way. Aim for "out of the way."
Written by the MetaMaxd team. We run Meta ads for brands that are tired of agency theater: no vanity metrics, no "hacks," just accounts engineered to make money. If your account looks like the cathedral described above, our audit will tell you exactly what to consolidate, and what to leave alone.
