Why Your Meta ROAS Is Lying to You (And How Post-Purchase Surveys Fix It)

Your Meta dashboard says 4.2x ROAS. Your Google dashboard says 3.8x. Your email platform claims 6.1x. Add them up and you're apparently generating 14x return on every dollar you spend. Sounds great. It's completely false.

This is the dirty secret of multi-touch digital advertising: every platform takes full credit for every sale it touched. Meta sees a click before the purchase and claims the conversion. Google did the same. Your email ESP too. The actual sale happened once — but it's been counted three times.

The sale happened once. Three platforms took the credit.

The Last-Click Illusion

Most brands default to last-click attribution, which sounds logical until you think about how people actually buy. A customer discovers your brand on TikTok, sees a retargeting ad on Instagram three days later, Googles your brand name, and then converts through a branded search ad. Who gets credit? Google's brand campaign — a channel that essentially captured intent you already created.

In practice, this inflates paid search ROAS by an average of 60-80% and deflates top-of-funnel channels like Meta and TikTok by the same margin. You end up cutting the channels that actually generate awareness and doubling down on the ones that merely harvest it.

Why Platform Data Can't Be Trusted

iOS 14 made things worse. When Apple limited tracking, Meta lost visibility into roughly 40% of conversions. Their solution — modeled conversions — fills in the gaps with statistical estimates. Meta is literally guessing and presenting those guesses as fact inside your Ads Manager.

Add ad blockers (now used by 43% of desktop users), cross-device journeys, and browser privacy restrictions, and you're left with attribution data that represents maybe half of what's actually happening.

The One Question That Cuts Through Everything

There's a method that predates digital advertising and still works: you ask.

After a customer completes a purchase, a simple survey question — "How did you hear about us?" — captures something no pixel ever can: the customer's own memory of their discovery moment. Not the last click. Not a modeled conversion. The actual channel that made them aware your product existed.

Picture a DTC coffee brand running this question post-purchase. Their Meta dashboard credits podcast ads with almost nothing, so they're about to cut the spend. But the survey tells a different story: a real share of buyers name a podcast as where they first heard about them — the discovery moment the pixel never captured. Cut that channel, and you'd be starving the top of your funnel without ever knowing it.

What Good Survey Data Looks Like

A well-implemented post-purchase survey running on a Shopify store will typically capture responses from 30-45% of buyers. That's enough statistical weight to make real budget decisions.

The key insight is that survey data and platform data should be used together, not in opposition. Platforms tell you what's converting. Surveys tell you what's driving awareness. When they disagree sharply — as they almost always do for top-of-funnel channels — the survey is usually closer to the truth.

Fixing Your Attribution Stack

You don't need to throw out your Meta campaigns. You need to stop letting Meta grade its own homework.

Set up a post-purchase survey on your confirmation page. Use a single question with 8-10 channel options relevant to your actual marketing mix. Let it run for 30 days. Then compare your survey-attributed channel distribution against your platform-reported distribution.

The gap you find is the distortion you've been optimizing around. Fix that, and your budget allocation decisions become grounded in reality.

rauxdata makes this setup take minutes, not weeks. Start free →