Fragmented accounts with dozens of narrow ad sets were built for the old Meta. Here is the consolidated structure that works with Andromeda, with example setups for small, medium and large budgets and a step-by-step migration plan.
Open many Meta ad accounts that were built before 2024 and you find the same pattern: campaigns for every audience, ad sets for every interest, age band and lookalike percentage, each with a small budget. That structure made sense when you had to tell Meta who to show ads to. With Meta Andromeda choosing from tens of millions of ads using what each ad contains, it mostly gets in the way.
This guide covers why consolidation works, the structures we use, and how to move an old account over without wrecking performance.
Why fragmented accounts struggle now
- Thin data per decision. Each ad set learns separately. Ten ad sets with ten conversions a week each learn less than one ad set with a hundred.
- Learning phase that never ends. Meta's ad sets need about 50 optimisation events in a week to exit the learning phase. Small ad sets rarely get there.
- You compete with yourself. Overlapping ad sets bid for the same people in the same auctions.
- Creative spreads too thin. With five ads in each of ten ad sets, no ad set has enough variety for retrieval to work with.
- Narrow eligibility. Every interest or lookalike limit shrinks the pool Andromeda can search for each ad.
The principles of an Andromeda-ready structure
- Split by goal, not by audience. Separate campaigns for different optimisation events, such as purchases versus leads, or for different products with different economics.
- Keep ad sets broad. Broad or Advantage+ audience, with location, legal age limits and exclusions as guardrails. See broad targeting on Meta.
- Fill each ad set with diverse creative. Different concepts, personas and formats in the same ad set. See creative diversification.
- Fund each ad set to exit learning. Aim for roughly 50 optimisation events a week per ad set. If you cannot, consolidate further or optimise for an event higher in the funnel.
- Use Advantage+ placements by default. Let Meta find the cheapest place to reach each person, unless brand safety or format rules say otherwise.
- Keep a place for testing. Either inside the main ad set or a separate testing campaign. See creative testing.
Example structures by budget
Small budget (under about 1,500 a month)
- One campaign: your main conversion goal, such as leads or purchases.
- One ad set: broad or Advantage+ audience, Advantage+ placements.
- Three to five ads: genuinely different concepts.
- Retargeting: none separately; prospecting already reaches warm visitors.
Medium budget (about 1,500 to 15,000 a month)
- Campaign 1, prospecting: one or two broad ad sets with five to ten concepts each.
- Campaign 2, testing: same targeting and event, 20 to 30% of budget, two to four new concepts every two weeks.
- Campaign 3, retargeting (optional): only if you have a large warm audience and a different message for it, such as pricing or proof.
Large budget (above about 15,000 a month)
- Advantage+ sales or leads campaigns for the core goal, with many diversified ads.
- A manual prospecting campaign for control and comparison.
- A testing campaign with a steady flow of concepts.
- Separate campaigns only for genuinely different products, countries with different economics or different optimisation events.
Self-assessment
Does your structure need consolidating?
Answer for your prospecting campaigns.
0 of 5 answered
How to migrate an old account safely
- Audit what you have. List every active campaign and ad set with spend, results and learning status for the last 30 days.
- Pick your goals. Decide the one to three optimisation events that matter. Everything else is a candidate to pause.
- Build the new broad ad set alongside the old ones. Use the best-performing concepts from across the account, keeping post IDs so social proof carries over.
- Shift budget gradually. Move about a quarter of the old budget to the new structure each week, rather than switching overnight.
- Compare on real outcomes. After two to three weeks, compare cost per real result, ideally checked in your CRM or revenue data.
- Pause the old ad sets once the new structure matches or beats them.
Expect some volatility in the first fortnight while new ad sets learn. That is normal. Do not judge on the first few days.
When splitting still makes sense
- Different optimisation events, such as a trial campaign and a demo campaign.
- Different products with different prices or margins.
- Different countries with different costs, languages or legal rules.
- Special ad categories like housing, employment, credit or politics, which have their own rules.
- Clear existing-customer offers that need exclusions or different messaging.
Quick quiz
Structure quick check
Three questions.
1. Roughly how many optimisation events per week does a Meta ad set need to exit learning?
2. Which is the best reason to create a separate campaign?
3. When moving to a consolidated structure, you should:
0 of 3 answered
Frequently asked questions
For small and many medium budgets, yes. It concentrates data and gives Andromeda the widest pool. Add campaigns only for different goals or business rules.
Both can work. Advantage+ campaigns automate more and suit accounts with good conversion data. Manual campaigns give more control. Many accounts run both and compare.
With few ad sets it matters less. With several, campaign budget lets Meta move spend to the ad set doing best, which usually helps.
New ad sets go through learning. Reusing post IDs keeps the engagement on your ads, and moving budget gradually limits the dip.
