As a statistician, it pains me to see marketers do the following:
โช Make million dollar marketing decision on just correlation
โช Specify Marketing Mix Models (MMM) without any statistical rigor
โช Think that MMM can be specified without controlling for all factors
โช Think confidence interval from disparate models and experiments can be compared
โช Think confidence interval talks about some sort of confidence
โช Poorly understand Randomized control Trials (RCTs) and think that it can be used to validate MMM
โช Think Beta coefficients of Multi Linear Regression models can be compared to ATE of experiments.
โช Poorly understand Design of Experiments and thereby conducting wrong experimentation (A/B testing, Geo testing etc.)
Now let me elaborate on the point of Statistical Rigor
What I like about the field of Biostatistics is that they follow statistical rigor and they must because our lives are literally on the line.
Such mission critical nature of a domain makes them apply statistical methods with a lot of care and nuance.
๐ Why do we need statistical rigor in MMM or Marketing in general?
I have seen posts from mmm vendors here on LinkedIn where they advocate laxity in following statistical rigor because well it is marketing and not exact science.
I vehemently disagree. I believe us marketers also need to imbibe the philosophy of statistical rigor in our MMM models and Experiments.
Why?
For the simple reason that real dollars are on the line and we owe it to our customers that they get accurate models (as much as possible) so that they don’t lose money.
Incorporating Statistical rigors makes our MMM models and Experiments more accurate. Most clients nowadays get statistics / AI. So it would be wrong on us to assume that they wouldn’t get it.
Even if they don’t, just like a good doctor explains a procedure to a layman, we must explain the statistics behind our methods to our MMM customers.
Aryma Labs is often mistaken to be a biopharma or CRO company. We are not a biopharma or CRO company. We are a Marketing Science company but one thing that we share with them is statistical rigor.
Also a small trivia: Did you know that the Hill function used for transforming media variables to capture diminishing effect has its origins in Biochemistry and pharmacology!!
The equation was formulated by Archibald Hill in 1910 to describe oxygen binding to Hemoglobin.
