How do you structure a Meta funnel for a high-ticket product?

Why does a considered purchase need three stages instead of one?

Nobody decides to spend $3,600 on a chicken coop off a single ad. A considered purchase at that price needs more than one contact with the brand before someone's ready to buy, so the account has to be built to earn a second and third contact, not just a first click. One campaign asking a cold stranger to buy immediately skips the part of the decision that actually needs to happen.

Compare that to an impulse purchase under $50, where one ad and one click can plausibly cover the whole decision. At $3,600, the buyer is going to research, compare, and hesitate before paying, whether the account plans for that or not. Building three stages instead of one is planning for the hesitation instead of getting surprised by it.

What does each stage actually do?

Warm-up engagement campaigns run first. Their job is to season the pixel: get it seeing people who show real interest, engagement, video views, page visits, so the platform has signal to learn from before a conversion campaign ever launches. Next, a purchase-seed conversion campaign runs, and its job is to build a real buyer signal off actual purchases rather than clicks or engagement. That's the harder, slower stage, because it has to prove out real buying behavior, not just interest. Once that buyer signal exists, a 1% lookalike retargeting campaign runs on top of it, compounding on everything the first two stages built. It isn't reaching a cold audience at all; it's reaching people who look like the buyers the account has already earned, which is exactly why its numbers run so much higher than the other two stages.

What did this look like on a real account?

Moveable Coops sells premium $3,600 chicken coops, and this exact structure is what ran on that account. Total spend was $2,433 against $87,686 in attributed revenue, a $101 cost per purchase, a 6.84% click-through rate, and a 2.01 frequency, meaning the ads weren't wearing out the audience even while running this hard. The lifetime ROAS was 36.0x. That figure sits on the third stage, the 1% lookalike retargeting layer, built on top of the warm-up and conversion work underneath it, not as an isolated result.

The weekly numbers show the shape of it better than the lifetime figure alone: spend climbed from $224 in the week of June 17 to $861 by the week of July 15, and ROAS moved 71.7x, 32.7x, 49.2x, 55.3x, 21.9x across those same five weeks. That's not a flat 36.0x every week. It's a range, and some of those weeks came in well below the lifetime average.

Why does the honest version of the number matter more than the impressive one?

My own answer to whether the 36.0x is real is the one on the homepage: "Yes, pulled directly from the Meta Marketing API, and it's the retargeting layer of a full-funnel system, which is exactly why it's high. I show the whole funnel's economics on the call, not just the prettiest campaign. That honesty is the point." A retargeting number without the two stages that built it is a badge, not an account. Showing where the number actually lives, and what it took to get there, is the difference between a case study and a cherry-picked screenshot.

This is also one account. The result can't be fully separated from the coop's own product-market fit, or from what was already true about that audience before the campaigns ran. One account is a signal that the structure works somewhere, not proof it's a repeatable formula for every high-ticket product.