How to Structure a Meta Ads Account for Scaling Past €10k/Month

Most Meta Ads accounts don't break because of bad creative. They break because testing and scaling are tangled into the same campaigns, so every budget increase disrupts the learning phase, and every new creative test cannibalizes a campaign that was already working. Once you separate the two properly, scaling gets a lot more predictable.

1. Split testing and scaling into separate campaigns

Keep one campaign (or a small set of ad sets) dedicated purely to testing new audiences, hooks, and angles on a small, fixed daily budget. Keep a separate campaign for scaling — this only runs creative and audiences that have already proven themselves in testing. Never let a scaling campaign's budget swings interfere with a live test, and never let a test's noisy early data influence a scaling decision.

2. Start broad, not hyper-segmented

Heavily segmented interest targeting made sense years ago. With Advantage+ and Meta's current machine learning, tightly split ad sets usually just split your budget and starve each one of the signal it needs to exit the learning phase. Start with broad targeting plus a few strong creative variations, and let the algorithm find the audience — your job shifts to feeding it clear creative signal, not manually carving out interest groups.

3. Separate prospecting from retargeting budget

Retargeting campaigns almost always show a better ROAS on paper, which tempts people to over-fund them. But retargeting only exists because prospecting is feeding it new visitors. Set a floor on prospecting budget (I usually keep it at 70-80% of total spend once an account is stable) so the top of the funnel never starves — otherwise your retargeting pool shrinks and the whole account's ROAS drops a few weeks later for no obvious reason.

4. Scale with budget steps, not big jumps

Increasing a campaign's budget by more than 20-30% at once resets enough of the delivery signal that Meta effectively re-enters a learning period, and performance dips right when you need it to hold. Scale in smaller, more frequent steps instead, and give each step 2-3 days to stabilize before the next one. It's slower, but it avoids the sawtooth ROAS pattern that spooks people into scaling back down too early.

5. Set a creative refresh cadence before fatigue hits

Frequency creeping past 2-3 on a given ad set is usually the first sign creative fatigue is about to show up in the numbers. Rather than reacting after CTR drops, build a refresh cadence into the account from day one — new hooks and formats entering the testing campaign on a set schedule, so there's always a next wave of creative ready before the current batch tires out.

Common mistakes I still see most often

• Fragmenting budget across too many ad sets, so none of them get enough volume to exit learning
• Judging a test on 2-3 days of data instead of waiting for a stable sample size
• Scaling budget and launching new creative in the same campaign at the same time — you lose the ability to tell which change caused what
• Letting retargeting audiences overlap heavily with prospecting, which just bids the account against itself

None of this is complicated in theory. What's hard is doing it consistently once real budget and real pressure to hit numbers enter the picture — that discipline is most of what a good account structure actually buys you.