Key takeaways
- A good Meta ads account structure in 2026 is consolidated: one main campaign per conversion goal, broad audiences, and many distinct creative concepts inside it.
- Split campaigns only when something real differs, such as the optimisation event, the country, the margin, or a budget you must protect.
- Keep creative testing in its own lane so new concepts get fair spend, then move winners into the main campaign.
- Structure only works on top of clean signal, so the Conversions API, deduplication and the right optimisation event come before any restructure.
The Meta ads account structure that works in 2026 is simple: a small number of consolidated campaigns, broad audiences, Advantage+ automation switched on, and most of your energy spent on creative. Each extra campaign or ad set you add splits the conversion signal Meta needs to learn. Split only when the business reason is real.
That is the short version. The rest of this article is how we set it up in practice, and where we break our own rule.
Why does consolidation win on Meta now?
Two things changed over the last few years.
First, Meta's delivery system got much better at finding buyers on its own. Its ad retrieval now leans heavily on the creative itself: who engages with a video, what the hook says, what the product looks like. Interest stacks matter far less than they did in 2020.
Second, the learning phase still runs on volume. Meta's guidance is still roughly 50 optimisation events per ad set in a week to exit learning. A clinic in Kemang generating 60 leads a week can feed one ad set comfortably. Split that into five ad sets by interest and each gets about 12. None of them learn. All of them look "unstable," so someone edits them, and learning restarts.
We see this pattern all the time. Fifteen campaigns, forty ad sets, and a CPA that swings every day. The fix is almost never a clever new audience. It is fewer containers.
What does a consolidated Meta ads account look like?
Here is the base structure we start most accounts with. Names vary by goal; the logic stays the same.
“On Meta, the creative is now the targeting. Structure just gets out of its way.”
| Campaign | Purpose | Audience | Budget | Creative |
|---|---|---|---|---|
| Main scaling | Drive the core conversion (purchase, lead, install) | Broad, Advantage+ audience | Campaign budget, most of total spend | 8 to 20 proven ads across several concepts |
| Creative testing | Give new concepts fair spend | Broad | Fixed ad set or test budget | 3 to 6 new concepts at a time |
| Special case (optional) | A different market, margin or event | Depends | Protected budget | Built for that case |
For ecommerce this main campaign is usually an Advantage+ sales campaign. For lead generation, an Advantage+ leads setup or a manual leads campaign with Advantage+ audience. For apps, Advantage+ app campaigns. Meta has been folding these into one streamlined creation flow, so the labels in Ads Manager may look slightly different in your account, but the principle holds.
Two settings we check on every main campaign:
- Existing customers. Define them with your customer list and site audiences so you can see how much delivery goes to people who already bought. If you sell repeat-purchase products, that might be fine. If you pay for new customers, watch it.
- Placements. Leave Advantage+ placements on unless you have evidence a placement brings junk. For lead forms, Audience Network sometimes does. Check by placement before excluding.
When should you split a Meta campaign?
Consolidation is a default, not a religion. We split when one of these is true:
- Different optimisation event. Registrations and purchases are different goals. They get different campaigns.
- Different country or currency. A Bali villa developer selling to Australia and Singapore buyers has different CPLs, languages and sales cycles in each. Separate them.
- Very different margins or values. A skincare brand with a hero serum and low-ticket sample kits should not let the algorithm chase cheap kit orders with serum budget.
- A budget that must be protected. A launch, a brand line the client insists on, or a market you are entering.
What does not justify a split: a new interest idea, a new age bracket, "men vs women," or a creative you want to try. Those go into the testing lane or the main campaign.
A pattern we see often: a local fashion brand in Bandung running separate campaigns for "women 18-24", "women 25-34", "hijab interest" and "office wear interest." All four chase the same purchase, for the same products, in the same market. That is one campaign cut into four. Merge them and let the creative sort out who responds.
How do you test creative without breaking the structure?
This is where most of the performance on Meta comes from, so it gets its own lane.
Test concepts, not tweaks. A concept is a different angle: a testimonial from a nurse, a price comparison, a founder talking to camera, a before and after. Swapping the button colour is a variation. Meta's system rewards creative diversity, so ten versions of the same idea behave like one ad.
Give tests a fair shot. Put new concepts in a testing campaign or use Meta's built-in creative testing feature, where each ad gets a controlled share of spend. In a big scaling campaign, a new ad often gets almost no delivery against proven winners, so it never gets judged at all.
Set graduation rules before you test. For example: after a set number of conversions or seven days, any concept within 20% of the main campaign's CPA moves across. Write the rule down so the decision is not mood-based.
Refresh on a rhythm. We aim to bring a few new concepts in every few weeks. Watch rising frequency and falling click-through on the top ads; that is fatigue showing up before CPA moves.
What signal does the structure need underneath?
A clean structure on top of bad data still learns the wrong thing.
- Conversions API alongside the pixel, with matching event_name and event_id so Meta deduplicates. We cover the full check in tracking audit before scaling ads.
- Event Match Quality. Send hashed email and phone server-side where you have consent.
- The right optimisation event. For lead gen, form submissions alone pull in junk. Send qualified leads back through CAPI and, where available, optimise toward them. For WhatsApp leads, qualify in chat before booking and pass the outcome back. Our click to WhatsApp ads guide goes deeper.
- Values. If you optimise for value, every purchase needs a real value and currency.
How we structure Meta accounts at FOXA
We audit first, then usually cut the number of campaigns and ad sets by more than half. Budget moves into one main campaign per goal, a testing lane gets built, and the signal layer gets fixed before anything scales.
For Pintarnya, a job app, that meant consolidating around registrations with a steady flow of new creative concepts. Meta cost per registration fell 68% and registrations grew 3.6x (the story). For My Prime Aesthetic, a clinic in Jakarta, we rebuilt the landing page, routed leads into WhatsApp and qualified them before booking. Cost per lead dropped 70% in six months (the story).
In both accounts the gains came from consolidation and a steady supply of new concepts, on top of cleaner data. No new audience trick was involved. If you want to see how Meta fits into the wider channel mix, read ROAS, CAC, CPA and LTV explained.
Want a second opinion on your account?
If your Meta account has more campaigns than you can explain in one sentence each, it is probably over-split. Book a call with Ads Manager open and we will point out which campaigns we would merge first.
Questions we get asked
How many campaigns should a Meta ads account have?
For most brands, two to four active campaigns is enough: one main scaling campaign per conversion goal, one creative testing campaign, and possibly one for a distinct market or product line. If you run more than that, check whether each campaign actually gets enough conversions per week to learn. If it does not, merge.
Is interest targeting still worth using on Meta in 2026?
Rarely as a hard restriction. Meta's delivery system finds buyers from broad audiences more reliably than most interest stacks, especially with Advantage+ audience switched on. Interests and custom audiences still help as suggestions, and for niche B2B or very local offers a tighter audience can make sense. Test it against broad rather than assuming.
How long should I leave a new Meta campaign before judging it?
Give it enough time to collect around 50 optimisation events in a week per ad set, or at least seven days if volume is lower. Judging after two days means you are reading noise. Avoid large edits during that window, because significant changes to budget, targeting or creative can restart learning.
Should I run separate retargeting campaigns on Meta?
Usually not as a standalone campaign anymore. Advantage+ campaigns already reach warm audiences, and a separate retargeting campaign often just bids against your own prospecting. Keep a small one only if you have a specific offer for past visitors or customers, and watch the overlap and frequency closely.
About FOXA DigitalFOXA Digital is a performance marketing agency in Jakarta, Indonesia. Since 2018 we have run paid media, tracking, SEO and content for apps, consumer brands, hospitality and attractions, judged on CAC, ROAS and leads.
Want us to look at your numbers?
30 minutes. No deck. You leave with three things to fix this week.



