Who's building TRANSPARENT attribution models? (No black-box BS)

This question is only relevant to people doing US$ 50K+ a month in advertising, across online AND offline channels. Skip if you're below that. it's not relevant to small budget brands.

Quick question for the group:

We all know Meta/Google attribution is basically a black box that's getting worse every month (thanks iOS 14.5+ 🙄).

While HYROS solves the MTA tracking piece brilliantly (!)

I'm curious: has anyone experimented with Marketing Mix Modeling (MMM) with FULL transparency? (No black-box solutions like so many platforms pretending to do MMM!)

I'm talking about approaches where you can actually see and audit every assumption in the model, and not just another vendor telling you to "trust the algorithm."

Specifically looking for:

Open source MMM frameworks that work and don't require you to be a software engineer or datascience PhD. Perhaps some you are building your own transparent models and want to share code.
Experiences with MMM approaches that are not snake-oil

The goal: finding an approach to honest attribution where we control the methodology, not Meta/Google controlling the data narrative.

I'm in the market for this if it exists.

Anyone going down this path? What's working? What's not?

While we're at it, does any of you also do "Triangulation" (Combining MMM/MTA/Econ/Holdout)?

Drop your experiences below so we can learn 👇

0 comments