MMM is a Game of Subjective Refraction
Couple of days ago, I went for my regular eye check up and the subsequent power correction.
If you wear glasses, you know the drill. The optometrist tries lens after lens, each one slightly different and asks you “Is the image clearer now or not?”
If you add too less or underpowered lens, the image looks blurred. But the moment you add a few more lens (with right power), the image magically becomes clearer.
This process is called subjective refraction.
📌 MMM parallel with Subjective Refraction
While the optometrist was trying many lenses and kept asking me “Better with this or that?”, My mind somehow saw similarities with subjective refraction and MMM.
Before MMM, most organizations are operating with a blurred view:
– Multiple teams claiming attribution for the realized KPI
– Channels appearing stronger than they are
– Overall decisions driven by partial visibility
📌 Enter MMM : Adding Lenses = Adding Variables
In the eye power determination, clarity improves not by chance, but through controlled iteration.
In MMM, it is the same.
Each variable you introduce is like a lens:
– The right additions sharpen the picture
– The system starts separating real drivers from background noise
The mirage (fuzzy attribution) starts turning into a clear image (clear attribution).
📌 Too many variables – clear image to blurred
If you keep adding lenses beyond what’s needed, the vision starts to strain, sharpness drops and the image blurs again.
In MMM too the same happens, too many variables and you find the metrics that really matters go off the rails.
📌 MMMDiagnose – Answering the ‘Is this clearer now or not’
In the eye test, you the patient provide the feedback on how well the image has become clear or blurry. For MMM, there are many calibration metrics that provide this feedback.
Our MMMDiagnose tool synthesizes nearly a dozen metric to provide one ‘Aryma Score’ which answers the question whether your model is providing a clear image of attribution or is it still blurred.
I will leave a link to our tool in the comments. It is free to try for up to 5 models and one just requires 3 inputs
1) Your ground truth (KPI)
2) Your predicted KPI
3) Number of variables you put in your MMM model.
📌 The Real skill isn’t adding more lenses or variables.
In the eye test one doesn’t aim for the maximum number of lenses.
The optometrist aims for the correct prescription (correct number of lenses but powerful ones)
MMM is no different.
It takes a lot of skill and domain knowledge to add just the right amount of variables.
Yes , much like the eye test, the MMM is also iterative.
It can’t be fully automated. But somethings are done well with careful, well informed iterations.
