Marketing Analytics Confidential: Get to Know MMM

We love #MMM (marketing mix modeling) around here. It’s multi-channel, it’s a great solution for post-cookie marketing, and accounts for incrementality, halo, marginal return, promotions and and you can even have it incorporate competitor behavior.  

That said, it’s not perfect, and it can point you in the wrong direction if you’re not keeping a critical eye on its input.

What do we mean?

First, you need at least two years of revenue and cost data to get meaningful insights. The more data volume, the better.

Second, whatever your volume, if all channel spend rises and falls evenly, MMM will have trouble finding incrementality trends – it needs media variance. The correlations it establishes need to be pressure-tested with on/offs after the initial model is made; it is more of a process than a one-time analysis. 

Third, MMM also struggles with long purchase cycles and does not incorporate customer lifetime value without deliberately altering revenue.  

So is it tricky? Yes, but the positives make it the most interesting and promising attribution tech that the industry has had in a long time.  

Stay tuned for our next Analytics Confidential deep dive!