Meta Ads Attribution: Why Your ROAS Is Lying to You

Your Meta Ads dashboard shows a 4.2× ROAS. Your blended MER (total revenue divided by total ad spend) is 2.1×. The gap between these two numbers is your attribution problem — and almost every eCommerce brand has it.

How Meta’s Attribution Window Works

By default, Meta attributes a conversion to an ad if the user clicked that ad within the last 7 days, or viewed that ad within the last 1 day, before converting. This means:

A customer who sees your Meta ad on Monday, ignores it, searches Google for your brand on Friday, and buys — that conversion is attributed to Meta’s 7-day click window. Google also claims it via last-click. Both platforms report the sale. You made one sale.

A customer who sees your Meta ad but never clicks, then buys direct on the same day — Meta claims that via the 1-day view attribution. It’s a sale that may well have happened without the ad doing anything at all.

The Incrementality Question

The right question to ask about any ad spend is: “Would these conversions have happened without this ad?” This is the incrementality question, and it’s the one your ROAS dashboard is specifically not designed to answer.

High-funnel brand awareness ads that reach people who would have bought anyway score well on ROAS but have near-zero incrementality. Lower-funnel retargeting ads that reach people in active consideration have high incrementality because the ad is actually influencing a decision that wasn’t yet made. The difference is enormous — and most brands can’t see it in their standard reporting.

How to Measure True Incrementality

• Meta’s Conversion Lift Studies: Creates a randomised holdout group — people eligible to see your ads but who won’t. Compares conversion rates between exposed and unexposed groups. The lift is your true incremental conversion rate. Run these for your main campaign types annually.
• Ghost Ad Tests: Run PSA (public service announcement) ads to a control group instead of your actual ads. More rigorous than Conversion Lift but requires more setup.
• Geo-Based Holdout Tests: Pause Meta ads in specific geographic markets for 4–6 weeks and compare revenue performance against comparable markets where Meta continues. Useful for measuring brand-level impact.
• Simple MER Comparison: Compare your blended MER against platform-reported ROAS. A large gap (e.g., 4× reported vs 2× blended) is a signal that significant over-attribution is occurring.

Practical Changes to Make Right Now

Change your attribution window: Switch from 7-day click, 1-day view to 7-day click only (or 1-day click for tighter measurement). This removes view-through attribution which is the largest driver of inflated ROAS. Reported ROAS will drop — this is correct, not a problem.

Compare Meta ROAS to blended MER weekly: Track both numbers together. If Meta ROAS is rising but MER is flat or declining, Meta is claiming credit for conversions it isn’t causing.

Segment retargeting from prospecting: Retargeting of warm audiences almost always shows high ROAS because these people are already considering buying. Don’t let high retargeting ROAS justify over-investment in prospecting that may be less efficient than reported.