You need new customers for your businesses?

Hellooo,

Everyone knows this move.

You're building a prospecting campaign. You want net-new people only. So you go into the ad set, open exclusions, and add every custom audience you've got.

All website visitors. Excluded. All purchasers. Excluded.

Clean. Done. Now Meta has no choice but to go find fresh people.

Except it does have a choice. And it's taking it.

Here's what your exclusion list actually contains

Most brands don't store their audience data independently. They're building exclusions out of Meta Pixel data.

And the Meta Pixel only retains website visitor data for the last 28 days.

Read that again if you need to.

Your "All Website Visitors" exclusion is not all website visitors. It's the last four weeks of website visitors. Everyone who browsed you in March, added to cart in April, bought in May, as far as that exclusion is concerned, they don't exist.

So they're eligible. They're warm. They're cheap to convert.

And Meta, being an obedient conversion-maximiser, walks straight toward them.

You built a wall with a 28-day gate in it and wondered why the same people kept walking through.

This is the exact reason your scaling breaks

You've seen this pattern. Everyone has.

The prospecting campaign performs beautifully at a low budget. ROAS holds. CPP looks sane. You get confident. You scale it 5x.

And it falls apart.

The standard diagnosis is creative fatigue, or auction pressure, or "Meta's being weird this month."

The actual diagnosis is simpler. At low spend, Meta only needs a small pool of the highest-probability converters and it's finding them inside your leaky exclusion window. At high spend, it needs volume, so it digs deeper into that same warm pool because that's still the cheapest conversion available to it.

The campaign didn't stop working. It was never built to find new people in the first place.

Quick pause — we're going through this live

We're running a session where we open real accounts and walk through the fix end to end: how the New Customer Purchase signal is configured, how the exclusion windows are set, and what the campaign structure looks like before and after.

Live dashboards. Real numbers. Questions answered on the spot.

👉 Save your seat for the upcoming session

Small room, on purpose. Bring your account.

The fix isn't a better exclusion. It's owning the list.

The moment your visitor and purchaser data lives in your own first-party layer instead of the pixel's memory, the 28-day ceiling disappears and you get to decide the window yourself.

Two audiences. Both yours.

All website visitors: set at 90 days, 180 days, or a full year, depending on who you genuinely don't want re-entering prospecting.

All purchasers: set to your actual definition of "new." If someone bought 6 months ago, is that a new customer to you? For some categories yes. For others you'd stretch it to 12. That's a business decision, and right now the pixel is making it for you at 28 days.

Then, and only then, does the exclusion do what you thought it was doing.

And while you're there, fix the event too

Exclusions control who sees the ad. The conversion event controls what Meta learns from.

Both need to be right.

New Customer Purchase: base event Purchase or Order Completed, filtered to Number of Orders = 1, from your source data. Not inferred. Not modelled. A fact.

Now the algorithm is being trained on first-time buyers only, and being kept away from everyone who already knows you. Those two things together are the acquisition engine. Either one alone is theatre.

One honest warning before you go build this

Give the ad set 5–7 days to learn, and expect fewer conversions than your account baseline at the start. Do not pause it on day 3. It normalises.

Also, this ad set is pulling new people into the top of the funnel, so a lot of the final conversion will land in your retargeting campaign. If you want to know whether it's actually working, turn it off and watch what happens to retargeting performance. That's your incrementality test, and it's free.

If you've never checked how far back your exclusion audiences actually go, that's the ten-second audit for this week.

Book a 1:1, we'll look at your exclusion windows together

Start your 14-day free trial

P.S. We have decided to give it for free this time. Better register, before we change our mind.