Beyond Marketing Drivers: A Conceptual Question #1303
Description
How should I think about Robyn MMM modelling for complex online SaaS business?
Let me setup a scenario.
As we know, in a typical SaaS there are key drivers beyond the traditional marketing, pricing, discounts, competitor and marcoeconomic data; for example: Companies like Xero or Intuit may have web funnel conversion optimisation program powered by A/B test which tries to understand how these changes can be used to make the customer (web traffic or visitors) journey smoother and get more conversions. Some of these traffic will sign up and register to the product by sharing this phone and email. Once they get their PII data they will (1) be called by the telesales team, or (2) get an onboarding email. Furthermore, those who sign up may experience another A/B test powered product feature release experiments. All these things are not marketing driver, nor control variables in traditional MMM; they are causal levers with real, measurable impact.
Q1: How should we think about using the Robyn framework to model such scenarios? Do you have an example implementation for such kind of model?
Q2: What is the implication of not including these data into the model? Assume we know these are initiatives which worked or did not work (and the magnitude of impact) because we have data from randomised controlled experiment or other means. Meaning, which subsequent decision afterfacts (coefficients, channel contributor, whatif scenario planner or budget planning) will be impacted and how?