Michael Kaminsky’s Post

View profile for Michael Kaminsky, graphic

I'm helping brands eliminate their wasted marketing spend 🔥. Follow me for essays on marketing effectiveness in an omni-channel world.

Should you control for website visits in your MMM? It's a great question, but the answer isn't straightforward. Naively controlling for website visits in an MMM can inject bias into your model because visits are on the path to conversion. Including them directly will skew your results and cause you to underestimate the impact of marketing on your business. If you want to model how website visits are impacting your sales, then the solution is to use a nested model. This approach allows you to estimate different parts of the funnel separately, correctly capturing the relationships between marketing spend, website traffic, and conversions. But here's the catch: it's complex. Really complex. You'll need to code the data-generating process by hand, estimate it carefully, and be prepared for longer computation times. There's no out-of-the-box solution for this level of complexity. Is it worth it? That depends on how critical understanding these relationships are for your business. If you’d like to dive deeper into the nuances of nested models and their implementation, check out the full clip for a more detailed breakdown.

Dr. Peter Cain

Executive Partner and co-founder at marketscience

3mo

True - although the web visit coefficient will be biased and needs to be identified.

Like
Reply

To view or add a comment, sign in

Explore topics