The Ecomm Analyst

The first thing I tell a new client looking at their Meta ads is to stop using the ROAS column as the answer. It’s data. It’s not the answer.

Meta’s reported ROAS is built on click-and-view-through attribution inside Meta’s own window. If someone sees an ad, doesn’t click, then buys two days later from an email, Meta will often still claim the sale. So will Klaviyo. So will Google. Add them up and you’ll find your platforms collectively claim 140% of your revenue.

This isn’t a scandal. It’s how the math works when every tool grades its own homework.

What I actually want to know is closer to “if I cut this campaign tomorrow, what happens to revenue?” Platform ROAS doesn’t answer that. It can’t. The platform has no idea what your email program is doing, what your influencer drops did last week, or whether you’re running a sale on a competitor channel.

I use ThoughtMetric as the layer that sits across all of it. Same orders, same window, same rules applied to every channel. When Meta tells me a campaign did $50K and ThoughtMetric tells me it did $18K on last-click and $24K data-driven, I now have a conversation to have. Maybe the campaign is still worth it. Maybe it isn’t. But I’m having that conversation with one source of truth, not five sources arguing.

The hard part for most operators isn’t choosing an attribution model. It’s getting to a place where every channel is being judged by the same yardstick. Until then, “ROAS” is a number that means something different depending on which tab you’re looking at.

Practical take. Keep Meta’s reported ROAS open. It tells you about delivery, creative fatigue, audience response, and pacing. Just don’t let it tell you which campaign to scale.