Original LinkedIn post
💡 Marketing Mix Modeling is only worth the effort if marketers are willing to take action based on the results
Since building my first Marketing Mix Model in 2012, I've seen and heard hundreds of marketers with different seniority levels and backgrounds reacting to MMM and its results. Here's when I've typically seen marketers take action based on MMM:
1️⃣ MMM results make sense to the marketer
If the marketer can't connect MMM results to their own reality and data, it is unlikely that action will follow. When the results contradict marketer’s intuition, there needs to be complementary data at hand to explain. For example, if there's a sudden change in ROI of a marketing activity, can it be explained by a change in media unit costs?
2️⃣ It's easy to get actionable recommendations that make sense
Historical results are great for learning, but planning the future is when the benefit of MMM is realized. MMM solution needs to help marketers in key decisions. If their company is behind the monthly/quarterly sales target, the marketer wants to understand how much and where to invest to get back on track. If they are planning next month's budget, they need to get recommended detailed weekly budget allocation.
3️⃣ MMM results have sufficient level of detail
As an example, MMM will have limited benefit to a performance marketing team if it only provides high-level ad platform comparisons (Google vs Meta). If they don't get campaign-level ROIs from MMM, they will continue using ROAS reported by ad platforms or attribution tool in campaign optimization.
Many marketers are of course interested in technical model validation as well, but demonstrating that the math is right won't help you if MMM fails in any of the three topics above.
Is there anything you would add to the list?