Digital marketing
Your ROAS is lying to you: attribution for ecommerce, minus the theology
June 12, 2026 · 7 min · ScaleLab
Ask Meta, Google and GA4 how many sales each drove last month, add them up, and you'll often exceed your actual revenue. Every system claims credit generously. Attribution isn't solvable perfectly — but it is manageable honestly.
01Why every number disagrees
Ad platforms count conversions within their own attribution windows and take credit for view-through conversions you might have gotten anyway. GA4's model distributes credit differently again. None of them is lying by its own rules — the rules just serve different masters.
The trap is picking whichever number flatters the current narrative. Fixed definitions, chosen once and kept, beat perfect models nobody trusts.
02MER: the number that can't lie
Marketing efficiency ratio — total revenue divided by total marketing spend — ignores attribution entirely. It won't tell you which channel worked, but it tells you the truth about the whole, every day.
Run MER as the executive metric and platform ROAS as a directional channel signal. When platform ROAS improves but MER doesn't, the platform is taking credit for demand it didn't create.
03Test incrementality when the stakes justify it
Geo holdouts and audience splits measure what actually happens when you turn spend off — the only real answer to 'would these sales have happened anyway'. Brand search campaigns and broad retargeting are the usual suspects worth testing first.
These tests cost revenue short-term, which is why almost nobody runs them — and why the ones who do stop wasting budget the others keep spending.
04A setup that fits on one dashboard
MER and contribution margin at the top. Platform-reported ROAS per channel below, labeled as directional signals. New-customer share tracked separately so retargeting can't dress up as growth.
That's attribution most brands actually need: honest enough to steer by, simple enough that everyone reads it the same way.