The 4-Campaign Meta Ads Structure We Use to Scale DTC Brands Past $100k/Month

A founder messaged me last month convinced his account was "perfectly structured" and still wondering why ROAS hadn't budged.

He'd split everything into 14 campaigns. Cold, warm, hot, lookalikes stacked five deep, interest tests siloed off on their own. It looked organised. It looked deliberate. And it had done almost nothing for the number he actually cared about.

Here's the thing about Facebook ads account structure for ecommerce: it's a control surface, not a lever. It decides where your spend lands and who it lands on. It doesn't manufacture a better return out of thin air. If your creative is weak, the cleanest structure in the world won't save you. But if your creative is working, a messy structure will quietly leak spend into the wrong audiences and you'll never know it happened.

So let me walk you through the exact structure we run for brands scaling past ~$100k/month. Four campaigns. Each one gets a single job, a rough budget share, and one exclusion rule that keeps it in its lane.

Why four campaigns and not fourteen

The instinct when things aren't working is to add more campaigns. More tests, more audiences, more "structure". I understand the appeal. It feels like control.

In reality it's the opposite. Every extra campaign splits your budget into a smaller pool, which means Meta gets less data per campaign, which means slower learning and noisier results. You end up managing complexity instead of managing performance.

The brands I see scale cleanly run lean. Four campaigns, each one doing one thing well. I'd rather have four campaigns I fully understand than fourteen I'm guessing at.

The concept that holds it all together is what I call swim lanes. Each campaign targets a distinct cohort, and exclusions stop those cohorts bleeding into each other. The point isn't perfect separation - water can flow between lanes a little, and that's fine. The point is that when you push budget into one campaign, you know exactly which type of customer that money is chasing. No swim lanes, no idea whether you're prospecting or just re-buying people who were going to purchase anyway.

That's the whole game with structure. It's not about the ROAS number. It's about knowing what your money is doing.

Campaign 1: Prospecting, the engine that finds new customers

This is the campaign that actually grows the business. Net new customers, people who've never heard of you.

Budget share: roughly 70-80% of total spend. If you're spending ~$60k/month, that's ~$42-48k living here. New customer acquisition is where scale comes from, so it gets the lion's share. Everything else is support.

The setup: one campaign, campaign-level budget, broad targeting. I'm not stacking ten lookalikes or siloing interests at this stage. Broad lets Meta's algorithm hunt across the whole eligible audience, and for most ecommerce brands that beats hand-picked interests once you've got a winning ad to feed it.

The exclusion rule: this is the important bit. Exclude your buyers and your warm audiences. For most accounts that means excluding anyone who's purchased in the last 180 days and any site visitor from the last 90 days. Those people belong to your other campaigns. If you let prospecting target them, you'll get flattering ROAS - because warm people convert cheaply - and you'll fool yourself into thinking your acquisition is healthy when really you're just retargeting in disguise.

I believe this is the single most common structural mistake I see in audits. Founders run "prospecting" with no exclusions, see a 2.4x return, and assume new-customer acquisition is fine. Strip the buyers out and the true cold ROAS is often closer to ~1.3x. That's the number that actually tells you whether you can scale.

One thing worth saying plainly: this exclusion doesn't make your ads perform better. It makes your reporting honest. And honest reporting is what lets you make the right call on budget.

Campaign 2: Retargeting, the safety net for warm traffic

People who visited, added to cart, started checkout, then drifted off. They already know you. This campaign nudges them back.

Budget share: around 10-15%. Retargeting audiences are small, so they don't need much. The mistake here is overfunding it - pour too much budget into a tiny warm pool and your frequency goes through the roof.

The setup: I'd build a couple of narrow audiences. Add-to-cart and initiate-checkout from the last 90 days as one tier, general site visitors from the last 30 days as another. Keep these genuinely narrow, not "narrow as a suggestion" that Meta quietly broadens back out to a million people. If your add-to-cart audience is only ~2,000 people, that's correct. That's the whole point.

The exclusion rule: manage this one by frequency, not by a hard budget. Watch your frequency column over a rolling 14-day window. You want it sitting around 5-7. If it climbs past 7-10, you're hammering the same people and it's time to ease off. Budget here is dynamic - you nudge it up when frequency is low and pull it back when it spikes.

A practical note: load plenty of creative variety into retargeting, not just your two top performers. Picture being shown the same two ads ten times in a week by a brand you half-remember. You'd tune out. Variety keeps a warm audience warm.

Does retargeting move your blended ROAS? A little, and it'll look great in-platform because warm people convert. But be honest with yourself - a chunk of these people were going to come back anyway. Retargeting catches the ones on the fence. It's a net, not an engine. Don't let a pretty retargeting ROAS talk you into starving prospecting.

Campaign 3: Retention, the campaign most brands skip

Existing customers. People who've already bought and could buy again. Most brands ignore this entirely on the paid side and leave it all to email, which is a mistake at scale.

Budget share: roughly 5-10%, and honestly only worth running once you've got ~1,000-plus past customers. Below that, the audiences are too thin to bother with - leave it to Klaviyo.

The setup: purchasers from the last 180 days as one audience, all-time purchasers (synced from your CRM) as another. The creative here has to be different. These people already know the product. Don't show them a "discover us" ad or tell a subscriber to subscribe again. Show them the new drop, the bundle, the reason to come back.

The exclusion rule: the two tiers exclude each other so you're not double-serving, and the creative does the rest of the work. Light-touch separation is enough here.

Here's my take on why this matters more than it looks. If your returning customer rate is sitting low - and for a lot of DTC brands it's under 20%, which to put it bluntly is barely better than one-off dropship buying - then every customer you acquire is a one-night stand. Retention won't fix a broken product or a weak email flow. But for brands with something genuinely re-orderable, a small, well-managed retention campaign quietly lifts LTV, and LTV is what lets you afford a higher CAC in prospecting. That's the connection people miss. Retention's real payoff shows up in what you can afford to spend up top.

Campaign 4: Scale, where your proven winners go to get fed

This is the misunderstood one. Scale isn't a place you create new ads. It's where your already-proven winners go to get more budget, deliberately.

Budget share: start it around 10% of the account and let it climb fast as winners prove out. On a mature account this can become one of your biggest line items.

The setup: here's the problem scale solves. In your prospecting campaign, when you find a winning ad and try to pour budget into that campaign, you can't control where the money goes. Meta might trickle it across ten other ads in the campaign. You wanted to back the winner and instead you fed the field.

So you run a separate scale campaign and "graduate" winners into it. When an ad proves itself in prospecting - top two or three by spend, comfortably above your ROAS target, and holding that trend for at least 2-3 days rather than one lucky day - you duplicate it into scale. You don't pause the original. It keeps running up top while its clone gets a bigger budget in a controlled environment.

The exclusion rule: same exclusions as prospecting - buyers and warm audiences out - because scale is still chasing new customers, just harder. Same swim lane, more force.

This is the part of structure that genuinely does move the number, and it's worth being clear about why. It's not the campaign that performs - it's the discipline. Graduation forces you to define what "winning" actually means before you spend hard behind an ad. Without it, your two biggest ads hog all the budget by default and nothing new ever gets a real chance to grow. A graduation rule is how you scale on purpose instead of by accident.

Where structure actually moves the number, and where it doesn't

Let me be straight about this, because it's the whole point of the post.

Structure does not create ROAS. I want to say that plainly because so much advice implies the right campaign layout is some hidden ROAS cheat code. It isn't. Your creative and your offer set the ceiling. Structure decides how much of your spend gets to operate under that ceiling instead of leaking out the sides.

Where it genuinely moves the number: clean exclusions stop you overpaying to "acquire" people you already own, so your real cold-acquisition efficiency improves. And a graduation rule means budget concentrates behind proven ads instead of being spread thin, so more of your spend runs at winning economics. Both are real. Both are about control, not magic.

Where it doesn't: no amount of restructuring rescues weak creative. I've watched accounts with genuinely messy structure outperform "textbook" ones by a wide margin purely because the ads were better. If your ROAS is stuck, restructuring is rarely the first fix. Look at your creative and your offer first.

So my honest advice: get the four lanes in place, get the exclusions right, get a graduation rule written down. Then stop touching the structure and go pour your energy into creative, because that's where the ceiling actually lives.

If you want a second set of eyes before you rebuild anything, that's exactly what a Signal/Noise Audit is for - we look at your account, your exclusions, your real cold ROAS, and your unit economics, and tell you where the biggest gains actually sit. Plenty of the time the answer isn't "restructure". It's something cheaper and faster.

What does your prospecting ROAS look like once you strip the buyers out? If you don't know yet, that's the first number I'd go and find.