Event-Level Attribution Discrepancies Explained

For example: Why Revenue and Profit Events (in rare cases) Don’t Match

If you are:

• Looking at a very short time frame, or
  
  

• Working from a limited data foundation (for example, by viewing individual ad level results), or
  
  

• Reviewing a very recent time period where, for example, the “Profit” event has already been attributed a conversion while the “Revenue” event has not yet been fully processed (server-side data can take up to 48+ hours to finalize),

The explanation can be found in the way marketing platform attribution works:

We are creating separate conversion events such as ReaktionRevenue and ReaktionProfit.
 When marketing platforms (e.g., Google Ads, Meta Ads) process and attribute this tracking data, they do so in the context of an isolated conversion event (e.g., only ReaktionRevenue).

Because of privacy measures and tracking limitations, these platforms often rely on modelled conversions to fill the gaps where user-level data is missing. Modelled conversions are based on statistical analysis and best estimates — which naturally introduces a degree of randomness into reporting.

This randomness is most visible when analyzing:

• Very short timeframes (e.g., daily reporting)
  
  

• Segments with limited conversions (e.g., at the individual ad level)
  
  

Additionally, if a customer interacted with more than one ad, the platform’s model might give one event (e.g., ReaktionRevenue) more credit to Ad A, while another event (e.g., ReaktionProfit) gets more credit to Ad B.

However, when looking at longer time periods and larger datasets, the impact of modelled randomness etc. usually evens out, and differences between conversion events tend to be much smaller.