If you add up the conversions reported by Google Ads, those reported by Meta, those reported by TikTok and those reported by LinkedIn Ads, you almost always end up with more conversions than you actually had. That is not an accounting error: it is duplicate attribution. Each platform claims conversions that other platforms also claim.
Understanding attribution properly is essential for making the right decisions about budget.
The attribution models that exist
Last-click: the conversion is attributed to the last channel before the purchase. It was the historical standard. It undervalues channels such as organic content and branding that generate demand but are rarely the last click.
First-click: the conversion is attributed to the first channel in the journey. It overvalues discovery but undervalues closure.
Linear: the conversion is shared equally across all channels in the journey. It is fair but does not reflect the reality that some channels are more decisive than others.
Time decay: channels closest to the conversion receive more weight. It’s reasonable but arbitrary.
Data-driven attribution (DDA): Google and Meta use ML to assign weight to each channel based on its statistical contribution to the conversion. It is the best model within each platform, but it only sees that platform’s data.
Why platform models aren’t enough
The fundamental problem: each platform only sees its part of the journey. Meta doesn’t know the user also saw Google ads. Google doesn’t know the user arrived from LinkedIn. Each platform claims more conversions than it actually generated.
To get a true picture, you need a unified attribution model that sees all channels simultaneously.
Marketing Mix Modeling: the industry’s response
El Marketing Mix Modeling (MMM) is making a strong comeback for two reasons: the deterioration of digital tracking (iOS 14.5, end of third-party cookies) and the availability of more accessible tools than traditional MMMs.
MMM uses aggregated data (spend per channel per week, sales per week, external factors such as seasonality) to estimate the incremental impact of each channel without needing individual user tracking. It is more qualitative than DDA but more honest.
Accessible MMM tools for SMEs: Robyn by Meta (open source), LightweightMMM by Google (open source), Mass Analytics, Recast.
For companies with a marketing budget of over 20.000€ per month, it is worth implementing at least a basic MMM.