Cross-Platform Attribution: Meta, Google, and the Full Picture

Meta says they drove 100 conversions this week. Google says they drove 90. Your backend shows 120 total conversions. The math doesn't add up because every platform claims credit for the same sales. Welcome to the cross-platform attribution problem—where each ad platform acts like it's the only one that matters, and the truth lives somewhere in between.

Multi-touch customer journeys are the norm. A user sees your Meta ad Monday, clicks your Google ad Wednesday, gets retargeted on Meta Thursday, then purchases Friday. Who gets credit? Each platform says "me." Building a unified view requires understanding how each platform attributes, where they overlap, and how to construct a more accurate picture.

Why Attribution Platforms Disagree

Each Platform Uses Last-Touch (For Their Touchpoints)

Meta credits conversions to the last Meta touchpoint before conversion. Google credits conversions to the last Google touchpoint. Neither sees the other's interactions, so each claims full credit for conversions they "touched" within their attribution window.

Different Attribution Windows

Meta's default is 7-day click, 1-day view. Google Ads defaults to 30-day click for Search, 30-day engaged-view for video. A conversion happening 10 days after a Meta click wouldn't be attributed to Meta but would be attributed to Google if they had a click in that window.

View-Through Inflation

Meta includes 1-day view-through attribution by default. If a user sees a Meta ad and converts within 24 hours without clicking, Meta claims it. Google's Display campaigns do similar. Both might claim credit for the same conversion based on an ad the user passively saw.

Cross-Device Blindness

User clicks a Meta ad on mobile, browses on tablet, purchases on desktop. If Meta can't connect these devices, they might miss the conversion. If Google caught the desktop search, they claim it. Reality: both platforms contributed.

The Overcounting Problem

Add up platform-reported conversions and you'll often exceed actual sales. This isn't fraud—it's each platform applying their attribution model independently:

This overcounting makes it impossible to calculate true ROAS by simply summing platform reports. You need a unified view.

Building a Unified Attribution View

Option 1: Single Source of Truth (Backend Data)

Use your e-commerce platform, CRM, or data warehouse as the single source of truth. Track actual sales with UTM parameters to attribute them to channels. This eliminates platform bias but loses granularity.

Implementation:

• Add consistent UTM parameters to all ad URLs (utm_source, utm_medium, utm_campaign)
• Capture UTMs in your analytics and tie them to purchases
• Attribute each sale to the UTM source that drove the converting session

Limitation: This is typically last-click attribution—the final touchpoint gets all credit. Multi-touch journeys still have attribution blind spots.

Option 2: Google Analytics 4 (GA4) as Hub

GA4 can receive data from multiple ad platforms and apply consistent attribution models across them. It's not perfect, but it's a neutral third party.

Implementation:

• Link Google Ads directly to GA4
• Import Meta, TikTok, and other platform conversions via manual upload or integrations
• Use GA4's "Model Comparison" to see how different attribution models change credit allocation

Limitation: GA4 still has blind spots for iOS users and ad-blocked visitors. Cross-device matching isn't perfect. But it's better than trusting each platform independently.

Option 3: Multi-Touch Attribution (MTA) Tools

Dedicated MTA tools (Triple Whale, Northbeam, Rockerbox, Measured) collect data from all platforms and apply unified attribution models. They attempt to track the full customer journey and allocate credit accordingly.

Benefits:

• Single dashboard across all channels
• Custom attribution models (linear, time-decay, position-based)
• Deduplication of cross-platform overcounting

Costs: $500-5000+/month depending on spend. Worth it for advertisers spending $50k+/month across multiple platforms. For smaller budgets, the cost may exceed the value.

Option 4: Marketing Mix Modeling (MMM)

MMM uses statistical analysis of aggregate data to estimate channel contribution. It doesn't rely on user-level tracking, making it privacy-compliant and immune to iOS restrictions.

Benefits:

• Works without cookies or user-level data
• Captures offline media (TV, radio, OOH) alongside digital
• Measures true incrementality, not just attribution

Limitations: Requires significant historical data (12+ months). Less granular—can't tell you which ad creative works best. Better for strategic budget allocation than tactical optimization.

Practical Cross-Platform Reconciliation

For most advertisers, a perfect unified view isn't feasible. Here's a practical approach:

Step 1: Establish Your Source of Truth

Pick one system as the source of truth for total conversions—usually your e-commerce backend or CRM. This number is non-negotiable. Platform-reported conversions must reconcile to this total.

Step 2: Calculate Platform Contribution Factors

If platforms report 230 conversions but you have 120 sales, the collective "inflation factor" is 1.92x. You can allocate this proportionally:

• Meta reported: 100 → Adjusted: 100/230 * 120 = 52
• Google reported: 90 → Adjusted: 90/230 * 120 = 47
• TikTok reported: 40 → Adjusted: 40/230 * 120 = 21

This is rough but gets you closer to reality than trusting raw platform numbers.

Step 3: Track Trends, Not Absolutes

Even with overcounting, platform trends are meaningful. If Meta's reported ROAS drops 30% week-over-week, something changed—even if the absolute ROAS number is inflated. Use platform data for trend analysis and relative performance comparison.

Step 4: Run Incrementality Tests

The gold standard for understanding true platform contribution is incrementality testing. Pause one platform in a geo-region while keeping others running. Measure the lift in that region vs. control. This reveals how much each platform actually contributes.

Platform-Specific Attribution Details

Meta Ads Attribution

Default: 7-day click, 1-day view. Includes modeled conversions for iOS users. View-through attribution can inflate numbers for high-impression campaigns. Click-through is more defensible.

Tip: Compare "7-day click only" to your default to see how much view-through is contributing. If 40% of conversions are view-through, question whether those would have happened anyway.

Google Ads Attribution

Default varies by campaign type. Search typically uses data-driven attribution across Google properties. Display/Video may use different windows. Enhanced conversions improve matching accuracy.

Tip: Google's "Attribution" reports (under Measurement) show how credit is distributed across campaigns. Use "Model Comparison" to see how last-click vs. data-driven affects your perceived winners.

TikTok Ads Attribution

Similar to Meta: 7-day click, 1-day view default. TikTok's events API (their CAPI equivalent) improves tracking. TikTok is more entertainment-focused, so view-through may be more legitimate—users watch rather than click.

Microsoft/Bing Ads Attribution

Often overlooked but valuable for B2B. Uses 90-day click attribution by default—much longer than Meta/Google. This captures B2B's longer sales cycles but can inflate numbers if compared to shorter-window platforms.

Common Cross-Platform Attribution Mistakes

Mistake 1: Summing Platform Conversions

Adding Meta + Google + TikTok conversions gives you a number higher than reality. Never present summed platform conversions as total performance—stakeholders will question why reported conversions exceed sales.

Mistake 2: Comparing Different Windows

Meta at 7-day click shows different numbers than Google at 30-day click. Before comparing, align attribution windows or acknowledge the comparison is apples to oranges.

Mistake 3: Ignoring Assist Value

The platform that gets last-click credit isn't necessarily the most valuable. A brand awareness campaign on Meta might introduce customers who later convert via Google Search. Meta "assisted" but Google gets credit. Use multi-touch reports to understand assist value.

Mistake 4: Over-Trusting View-Through

View-through conversions are real, but they're less certain. A user who clicked your ad demonstrated intent. A user who saw your ad in their feed might have converted regardless. Weight click-through more heavily in cross-platform analysis.

Mistake 5: Siloed Optimization

Optimizing each platform in isolation can lead to suboptimal total results. Cutting Meta spend might reduce the awareness that feeds Google Search conversions. Think holistically about the customer journey.

The Customer Journey Reality

A typical customer journey looks like this:

1. Awareness: Sees Meta video ad (doesn't click)
2. Consideration: Clicks Google Search ad, browses site
3. Retargeting: Sees Meta retargeting ad (doesn't click)
4. Conversion: Direct types URL and purchases

In this scenario: Meta claims view-through credit. Google claims click credit. Direct gets the sale in GA. All three are partially right and wrong.

Understanding that most conversions are multi-touch helps you make better budget decisions than any single platform's attribution model.

Practical Budget Allocation Framework

Start with Platform-Reported Efficiency

Use platform ROAS as a starting point. Even with overcounting, relative efficiency (Meta ROAS 3x vs. Google ROAS 2x) gives directional guidance.

Apply Incrementality Discounts

Based on incrementality tests (or estimates), discount platform-reported conversions. If incrementality testing shows Meta's true lift is 70% of reported, adjust your ROAS calculation accordingly.

Consider Upper-Funnel Value

Platforms that drive awareness (Meta, TikTok video) may have lower last-click ROAS but create the pipeline that lower-funnel platforms (Google Search) convert. Don't starve upper-funnel just because it doesn't get last-click credit.

Budget to Tests

Reserve 10-20% of budget for incrementality tests. Regularly pause platforms in test regions to validate their true contribution. This data is more valuable than any attribution model.

Key Takeaways

• Every platform claims credit for multi-touch conversions—expect overcounting
• Use backend sales data as your source of truth for total conversions
• GA4 or MTA tools provide a more neutral cross-platform view
• Incrementality testing is the gold standard for understanding true platform value
• Compare platforms using aligned attribution windows
• Upper-funnel platforms provide value that last-click attribution misses
• Optimize holistically—cutting one platform may hurt others

FAQ

Which platform should I trust for attribution?

None of them completely. Trust your backend for total conversions. Use platform data for trends, relative performance, and optimization signals. Use incrementality tests for budget allocation decisions.

Is 90% overcounting normal?

For advertisers running multiple platforms with overlapping audiences, 50-100%+ overcounting is common. Advertisers with distinct audiences on each platform (e.g., Meta for B2C, LinkedIn for B2B) see less overlap.

Should I use the same attribution window on all platforms?

For fair comparison, yes. But each platform has different defaults and some don't offer the same options. Where possible, align to 7-day click for consistency. Document differences when you can't align.

How often should I run incrementality tests?

Quarterly for major platforms. More frequently if you're making significant budget changes or launching new channels. The data decays as market conditions change, so regular testing is necessary.

Are MTA tools worth the cost?

At $50k+/month in ad spend across 3+ platforms, usually yes. The cost of misallocated budget exceeds tool costs. Below $20k/month total spend, manual reconciliation and GA4 may be sufficient.