Back to stack
Layer 05/ 05

Measurement.

Server-side tracking and attribution, without leaking data to third parties.

Since the iOS updates and cookie banners, client-side tracking only shows you a slice of your conversions. Statistical MMM works on aggregate data instead, with no PII or cookies involved, and still makes channel contribution and saturation curves measurable.

Bayesian
MMM on aggregate data
Weekly
model refresh
0 PII
privacy-native
Curves
saturation per channel
How I solve it

How I set it up for you.

  • 01The model works on totals, not on individual people — it needs no personal data.
  • 02For each channel you see where the next euro stops paying off.
  • 03Recrunched once a week, written up for the board once a month.
Toolchain
PyMC-MarketingBigQueryAirbyteCloud RunLooker StudiodbtGeo TestsBayesian InferencePyMC-MarketingBigQueryAirbyteCloud RunLooker StudiodbtGeo TestsBayesian Inference
Saturation curves

Each curve = one channel. Y-axis: contribution to revenue. X-axis: spend. Flat zone = diminishing returns, that's where the money burns.

Example workflow

Example: Weekly model refresh

  1. 01Channel spend syncs into BigQuery daily (Fivetran/Airbyte)
  2. 02The PyMC-Marketing job runs weekly on Cloud Run
  3. 03It spits out saturation curves and the contribution plot
  4. 04C-level opens the Looker dashboard and reallocates the budget
05

Want me to build this for you?

30 min demo. I walk you through a real setup, live.