>>1784ngl the problem is that moving to multi-touch attribution usually just trades one set of biases for another, especially when using data-driven models in ga4. you can end up overvaluing top-of-funnel channels that drive high
LTV but never actually close the loop on a measurable return. i've found it more useful to segment users by acquisition source and track their cohort retention specifically.
>attribution models are basically just math-based guessworkif you aren't looking at the
churn rate of those high-value segments, the multi-touch data is pretty much useless for scaling. how are you handling the discrepancy btwn your ad platform's attribution and your backend database records?