Run this exercise once and you'll never trust a platform dashboard the same way: take last month's conversions as reported by Meta, Google, and your email tool. Add them up. Compare against actual sales from your own database. In most accounts the platforms collectively claim 2–3× your real number.
Nobody is technically lying. Each platform counts a conversion if it touched the buyer within its attribution window — a view, a click, an open. A customer who saw an Instagram ad, later googled you, clicked a search ad, and finally bought from an email link is one sale to you and three conversions to them. Every platform grades its own homework, and every platform passes.
The three layers of honest measurement
Layer 1 — Platform metrics, demoted to operating signals. Keep them, but reassign their job. CTR, CPM, frequency, platform-reported CPA: these are fine for within-channel decisions (“creative B beats creative A”). They are not evidence that the channel is creating customers. Treat them like a car's tachometer — useful for driving, useless for navigation.
Layer 2 — Your own conversion record. One pixel or server-side event on your property, writing to your database, joined to UTMs you control. This is the single source of truth that platform numbers must reconcile against. It's an afternoon of engineering and it permanently changes the quality of every argument your team has.
Layer 3 — Incrementality. The only question that matters is the counterfactual: how many of these sales would have happened anyway? Branded search ads famously “convert” people who were already typing your name. The honest tools here are holdout groups (deliberately exclude a slice of the audience and compare), geo splits, and difference-in-differences when you change spend in one place and not another. None of this requires a data-science team — it requires deciding, in advance, that you'd rather know.
A worked example
Say Meta reports 120 conversions at a $42 CPA and you're deciding whether to scale. Layer 2 says your database recorded 70 sales with Meta-tagged UTMs — already a different picture. Layer 3: you held out 20% of the audience last month, and the held-out group converted at 80% the rate of the exposed group. So Meta's true incremental contribution is roughly 70 × (1 − 0.8) ≈ 14 sales, putting real incremental CPA near $360, not $42. Maybe that's still worth it for your LTV. But now you're deciding with the truth.
Where Mayaa fits
This stack is wired into Mayaa rather than bolted on: every client gets first-party conversion tracking joined to channel activity, every paid campaign launches with a holdout by default, and our analytics agent reconciles the three layers before any “scale it” recommendation reaches your approval inbox. We'd rather report 14 true conversions than 120 flattering ones — because the next budget decision is only as good as the number it stands on.
Want this run for you? The team that wrote these posts is the team that does the work.
Start free trial
