Causality is the cheat code for accurate predictions

In a recent client project, we were pleasantly surprised to see that our model had predicted net new customer revenue with 95% accuracy for the future month of March 2026.

Generally Marketing Mix Modeling (MMM) don’t achieve this much high accuracy in the future periods. But Media Mix Modeling (MeMM) can.

If you are a D2C brand operating only with Meta ads or Google Ads or TikTok or just any of the two, it is easier to predict KPI accurately.

📌 How is this possible?

Well firstly MMM is implicitly causal in nature. But if we control for all variables and OVB problem, then your MMM can turn causal. Of course the validity of causality still needs to be confirmed via tests like DID. But one of the tell tale signs that your MMM is turning causal (assuming you have well specified model) is that it can predict well in the future periods.

📌 MeMM is easier to turn causal than MMM

Media Mix Modeling (MeMM) is far easier to turn causal as they have fewer variables and we can almost account for all of it and there is generally less OVB problem.

On the other hand, Marketing Mix Modeling (MMM) is more difficult to turn causal as there are so many more variables, there could be problems of OVB.

📌 Why MMM Struggles (and it should)

Full MMM operates in a much messier world:

– Pricing changes
– Distribution shifts
– Competition dynamics
– Adstock effect
– Interaction effects of many variables

You are essentially trying to model a complex and perhaps partially observed system.

So yes, prediction becomes harder because turning MMM to be causal is also equally harder.

But here’s the important nuance, lower predictive accuracy in MMM is often a sign of realism not weakness.

📌 Causality is hard

Establishing causality is hard. Especially in marketing, where we can’t perform RCTs. We have to rely on Quasi Causal methods.

If your model predicts well without being causal, it is mostly luck.

If it predicts well because it is causal, that is true leverage.

As I say in our causality course – causality makes strategy repeatable.