If you’ve just found out that marketing mix modelling can’t actually measure advertising’s full ROI then don’t panic. Econometricians know that already and most will freely admit it. It’s not a secret.
Nothing else can measure full ROI either.
The most useful thing you can do with that knowledge is to avoid the ROI measurement trap.
Anyone who’s watched marketing mix modelling projects being debriefed to finance directors will have seen the ROI measurement trap. You fall into it when you say that you say that you’ve commissioned MMM to measure advertising’s true ROI, present the number and the FD calculates that it’s not big enough to be profitable.
Now you have to explain that your project doesn’t really measure true ROI, it only measures the short to medium term part. There might also be a long term part that we can’t measure. Only we’re not sure how big the long term part might be. Your budget for next year is looking shaky…
It’s a trap. Don’t fall into it.
Use econometrics for what it’s good at. Play with different campaign scenarios and choose the best one, predict sales for the next quarter, simulate price changes. But don’t expect it to capture the full value of advertising because it can’t do that.
What can you do instead to understand the long term?
Let’s answer that question with a question. How did you know that Ullswater would dry up if it never rained again? The model didn’t understand that but you did.
You knew because you understand at least the basics of how lakes work and where the water comes from. You got that from intuition and you got it from experience – Ullswater hasn’t dried up but other lakes have.
We can do the same for our sales scenarios. We need benchmarks and examples from brands that did stop advertising. We need large cross-industry studies because all of the answers aren’t in your own data. And we need a mental model of how we believe advertising actually works. One that we can’t get from a marketing mix model.
If you’d like to read more (perhaps after a break and a cup of tea), I’ve collaborated with two colleagues to produce a guide to integrating marketing mix models with other techniques.
My company, Sequence Analytics, exists to weave together different advertising measurement tools and techniques because no one tool can do everything on its own. Not even marketing mix modelling.