There can be little doubt that marketing mix modeling (MMM) is having a moment. With the death of the cookie and increasing data privacy concerns from consumers, marketers are seeking alternatives to degrading multi-touch attribution (MTA) systems. However, as this moment unfolds, I’ve realized something critical: Understanding of marketing mix modeling solutions is woefully out of date. The industry has been stuck in a time warp, and marketers need to understand what’s changed to avoid settling for less.
Beyond spreadsheets: A new approach to marketing insights
Traditional marketing mix modeling has been failing businesses. Working with brands like Domino’s and Samsung, I’ve seen firsthand how the old consultant-driven approach with six-month result lags isn’t empowering businesses to make good decisions. The marketing mix modeling of yesterday was a clunky tool that left companies swimming in data but starving for insights.
If you’re looking at the marketing mix landscape today, you should be assessing key characteristics that will enable the transition from insight to action and results: modeling methodology, granularity, data handling, accuracy, transparency and support. You’re no longer looking for a spreadsheet; you’re looking for a growth co-pilot.
The death of bespoke models: Embracing flexibility
Let’s break down why the old ways are dying. Modeling methodology has fundamentally transformed. Where regression once dominated, modern providers now use Bayesian techniques. But the evolution goes deeper than the mathematical approach to modeling.
The future of modeling is generalized and time-varying, not specific to one business at one point in time. Generalized models can change with a business and its context, making them far less vulnerable to overfitting. Overfitting is where one variable correlates highly within a data set, but the observations have no real causal relevance, which is a high risk, especially when you have less observations that exist only in one specific context. Bespoke models that need to be rebuilt for every model run are a relic of the past.
Granularity: The key to actionable insights
Granularity is where the real modeling magic happens. Marketers need high-level numbers for C-suite discussions, but they also need performance insights across geographies, campaigns and time frames. They need to assess the impact of pricing, promotions and brand equity. The days of static PowerPoint presentations where data can’t be interrogated beyond a topline result are over.
From months to minutes: The data revolution
Data preparation, once a months-long nightmare, can now be accomplished in minutes. Every new generation MMM provider will offer automation via API for digital channels. But what about everything else? Modern marketers need solutions that enable them to clean and structure data from linear TV and out-of-home channels with just a few clicks as well.
Transparency: No longer optional
How much can you truly rely on your model? The gold standard is holdout testing. If your vendor isn’t conducting and sharing these tests, it’s time to ask why.
AI: Not just for creatives
Here’s something that might surprise you: Generative AI isn’t just for creative departments. Other companies in the mar tech stack (think Salesforce or Microsoft) have built generative AI tools to help customers query their models instantly, from media plan adjustments to board report preparations.
Marketing mix modeling has evolved from a necessary evil to an exciting opportunity for businesses to truly understand their marketing impact. But businesses today can’t settle for a PowerPoint presentation delivered in six months.
When we pioneered our next-generation marketing mix modeling methodology, we weren’t just creating another software solution; we were reimagining how businesses understand their marketing impact. That’s why we connect marketing efforts to financial outcomes. Our actionable insights that optimize ROI are how we differentiate our tech from yesterday’s MMMs.
There has never been a better time to explore next-generation solutions. The future of marketing measurement is more powerful, transparent and accessible than ever before.
For more articles featuring Henry Innis, click here.
