Marketing Modeling

For most companies, making the decision to invest in modeling such as MMM, MTA or similar, often comes with significant expectations.

These expectations are a core selling point for most companies that offer these services.

As a result, I have observed that there is a significant amount of scrutiny and analysis that goes into the tire-kicking process when the results and recommendations are presented back to the business.

With MMM as an example, this scrutiny often leads to over-fitting of models, as this helps in the tire-kicking process.

The real challenge with over fitting models is that they rarely hold up under a real live scenario and, consequently, your forecasting ends up being inaccurate.

The real expectation should be that your initial model is just the beginning of your data-driven journey.

Once you have your first model, the next phase should be to validating it by testing its ability to provide insights which can drive real results. This allows you to refine the model and establish trust in the process, which I believe should be the real goal.

From here, you can continue to build trust and start to deliver on the promised value of data modeling.

#marketinganalytics #marketingdata #datamodel