Your Meta Ads account shows 8x ROAS. But when you cross-reference with GA4, you see only 3x ROAS for the same campaigns. The discrepancy is almost entirely explained by Meta’s attribution window settings. Meta counts conversions differently based on whether users clicked or viewed your ad, and over what time window. This guide explains how Meta’s attribution windows work, how they inflate reported ROAS, and which settings to use for accurate measurement.

What Meta’s Attribution Windows Mean

Meta attributes a conversion to your ad if a user clicked your ad within the click attribution window (1-day, 7-day, or 28-day) OR viewed your ad within the view-through attribution window (1-day). Default setting: 7-day click + 1-day view. This means Meta takes credit for any purchase that happened within 7 days after a click OR 1 day after viewing your ad — even if the user bought from Google Search in between.

How Different Windows Change ROAS

Example: A customer sees your Meta ad Monday, then buys Saturday after clicking a Google Search ad Friday. With 7-day click window: Meta takes credit (purchase was within 7 days). With 1-day click window: Meta doesn’t take credit (5 days after click). With 1-day view included: even if they only saw the ad and never clicked, Meta takes credit if they buy Tuesday.

Recommended Settings by Use Case

• E-commerce: 7-day click only (turn off 1-day view) — most purchases happen within a week
• B2B / Long Sales Cycles: 28-day click or 7-day click — decision cycles are longer
• Cross-channel comparison with GA4: 1-day click only — closest equivalent to GA4’s last-click model

// Changing attribution window in Meta Ads Manager:
// 1. Go to Ads Manager → Columns → Attribution Setting
// 2. Select "1-day click" or "7-day click"
// 3. Uncheck "1-day view" to remove view-through
// 4. Use "Compare Windows" feature to see how each window changes numbers

Reconciling Meta ROAS with GA4

• In Meta Ads: set attribution window to “1-day click” (no view-through)
• In GA4: check “Paid Social” channel in Acquisition → Source/Medium (facebook / cpc)
• Expected variance: 10-20% (Meta may count some conversions GA4 doesn’t due to iOS)

Meta’s Aggregated Event Measurement (AEM)

Post-iOS 14, Meta uses AEM for iOS users, reporting modeled (estimated) conversions for opted-out users. These inflate conversion counts beyond what GA4 can measure. GA4 sees 0 for iOS opted-out users; Meta estimates their conversions statistically. This gap is partially unavoidable — focus on incrementality testing to understand Meta’s true causal impact.

Related guides: Meta Conversions API Field Mapping, Meta Pixel Deduplication, TikTok vs GA4 Reconciliation.