Conversion Deduplication: Meta, Google, TikTok | Metrikia
Ad tracking and analytics
Tracking & Attribution4 minFeb 24, 2026Updated Aug 7, 2026
BN

Baptiste Noel

Growth and co-founder of Metrikia

  • Master en neurosciences et neuropsychologies cliniques
  • Master en entraînement et optimisation de la performance
  • Créateur SaaS et de contenu, 20 000+ abonnés LinkedIn

Co-founder of Metrikia, Baptiste is building a SaaS from scratch and shares the growth journey unfiltered. A former clinical-neuroscience researcher and physical-performance coach, he built then left a coaching business generating over 70,000 EUR per month before focusing on product. He writes about growth strategy, acquisition and scaling.

LinkedIn

Conversion Deduplication Across Meta, Google, and TikTok: Stop Double Counting

Meta, Google, and TikTok each claim 100% of your conversions. Learn how Metrikia deduplicates conversions using CRM data and 9 attribution models.

Partager

The Double-Counting Problem: 1 Sale, 3 Platforms Claiming Credit

Picture this: a prospect sees your Instagram ad, clicks a Google ad two days later, then completes a purchase after clicking a TikTok ad. Here is what your dashboards show:

  • Meta claims the conversion (view + 7-day window)
  • Google claims the conversion (click + 30-day window)
  • TikTok claims the conversion (click + 7-day window)

Your CRM records 1 sale at 500 euros. Your platforms count 3, totaling 1,500 euros in reported revenue. Your real ROAS is divided by 3 compared to what platforms display.

This is not a bug, it is the platforms' business model. Each one has an incentive to claim as many conversions as possible to justify your ad spend.

The Scale of the Problem

On a 10,000 euro budget spread across 3 platforms, it is common to see:

  • Meta reports 25,000 euros in generated revenue
  • Google reports 18,000 euros in generated revenue
  • TikTok reports 12,000 euros in generated revenue
  • Total reported: 55,000 euros in revenue
  • Actual CRM: 25,000 euros in real revenue

The gap is massive. And many agencies base their optimizations on these inflated figures.

How Each Platform Handles Internal Deduplication

Meta: The omni_purchase Priority System

Meta uses a priority system to avoid internal double counting (between Pixel and CAPI). The hierarchy is:

  1. omni_purchase: unified cross-channel event (highest priority)
  2. fb_pixel_purchase: purchase detected by Facebook Pixel
  3. purchase: standard event sent via CAPI

Metrikia leverages this hierarchy: when multiple events exist for the same conversion, only the highest-priority event is retained. This prevents Pixel + CAPI double counting, a trap many advertisers fall into.

Google: Enhanced Conversions and GCLID

Google uses the GCLID (Google Click ID) to trace the path from click to conversion. With Enhanced Conversions, Google matches on first-party data (hashed email) to improve post-iOS 14 measurement.

The problem: Google only deduplicates internally. If the same user also clicked a Meta ad, Google does not know and claims the conversion anyway.

TikTok: click_id and Event Deduplication

TikTok uses a click_id for post-click tracking. Deduplication happens on the click identifier, if two conversion events share the same click_id, only one is counted. But like Meta and Google, TikTok knows nothing about conversions attributed by other platforms.

The Metrikia Solution: CRM as the Single Source of Truth

Cross-platform deduplication cannot come from the platforms themselves. It must come from an independent source of truth: your CRM.

The Principle

  1. Platforms report touchpoints: every click, every view, every interaction is recorded as a touchpoint in Metrikia
  2. The CRM records real conversions: a lead that signs, a deal that closes, that is the business truth
  3. Attribution distributes credit: instead of letting each platform claim 100%, Metrikia's 9 attribution models distribute credit proportionally

The 9 Attribution Models

Metrikia offers 9 models to distribute conversion credit:

  • First Touch: 100% to the first touchpoint
  • Last Touch: 100% to the last touchpoint
  • Linear: equal credit to each touchpoint
  • Time Decay: more credit to recent touchpoints
  • U-Shaped: 40% first + 40% last + 20% split across the middle
  • W-Shaped: first, middle, and last touchpoints weighted higher
  • Full Path: first, lead creation, opportunity, last, each weighted higher
  • Markov Chain: probabilistic model based on channel transition paths
  • Shapley Value: game theory, marginal contribution of each channel

Each model uses the CreditDistributor with the Largest Remainder Method to guarantee that credited amounts sum to exactly the conversion value. No rounding that creates phantom cents.

Concrete Example

A 5,000 euro deal with 3 touchpoints: Meta (view), Google (click), TikTok (click).

With the Time Decay model:

  • TikTok (most recent): 2,500 euros credit
  • Google (intermediate): 1,500 euros credit
  • Meta (oldest): 1,000 euros credit
  • Total: 5,000 euros (exactly the CRM value)

Instead of 3 x 5,000 = 15,000 euros claimed by platforms.

Impact on Your Decisions

When you switch from platform figures to deduplicated figures, decisions change:

  • The channel you thought was most profitable may not be
  • A channel underestimated in last-click may be crucial in first-touch
  • Budgets get reallocated toward the real conversion drivers

Deduplication is not a technical detail. It is the difference between flying blind and piloting with reliable data.

Try Metrikia free for 14 days and discover your real deduplicated ROAS.

To see how this piece fits the whole, see our guide to building a complete tracking system, layer by layer.

Ready to measure your true advertising ROI?

Connect your ad accounts and CRM in 5 minutes.