What to expect (and not expect) from MMM?
Marketing mix modeling is a macro analysis at the budget level. It does not rely on visibility into the consumer's journey, or user-level data such as impressions and clicks needed to run the model. MMM analysis did not yield any campaign-level strategy or creative that was performing better than the others. This has nothing to do with multi-touch or last-touch behavior at the user path level. It is a way to approve or reject decisions made at the budget level.
Cost issues and data accuracy are the most common limitations of marketing mix modeling. But much of this concern stems from preconceived beliefs that cloud the true value of the stock.
Here's what brands need to know about the four biggest MMM myths.
Myth no. Myth #1: The marketing mix model tells you what to do with your budget.
The media mix model can optimize budget allocation across channels based on historical data. For example, a brand may be 15% overfunded for TV spending and 5% underfunded for paid social advertising, and the breakdown of that breakdown by region between the West and East Coasts.
This information cannot predict what budgeting strategies and tactics may or may not work in the future. MMM results can be used to create predictive models, but they are not predictable. A brand may wish to introduce new channels that were not in the original model, or its business objectives may switch to other priorities where the calculated mix of channels is no longer relevant.
Myth no. Myth #2. Building a sustainable business is expensive for an organization.
Industry best practice guidelines often lead marketers to believe that building MMM functionality requires six-figure investments in enterprise-grade data warehouses and terabytes of historical data.
This is the truth for most brands. If you have historical data about your spending (this can be easily transferred from platforms like Google Ads, Meta and Pinterest) and your conversion metrics, you can model a marketing mix. Relatively low investment.
Myth #3: If I can't measure everything, I shouldn't measure anything.
Most brands don't need to boil oceans to derive value from MMM research. Instead of looking at weekly product sales or exit market strength for every competitive brand that impacts your industry, just look at the macro level how much money is spent on media each week for each channel and cover sales data. Regression models with simple data sets can produce mixed baselines that reject or support previous budget decisions.
Myth no. Myth #4. The AI powered platform is the answer to all my problems.
Advances in machine learning and artificial intelligence will bring significant benefits to marketing mix modeling. But AI alone cannot solve all MMM problems.
Depending on the question being asked and the database, some methods may work better than others. When selecting third-party tools and vendors, marketers must determine which method is best for their data and objectives, and understand what parameters in the model need to be adjusted to ensure the most relevant and comprehensive conditions possible.
A new kind of data addiction
As marketers continue to evaluate the role of marketing mix modeling in their organizations, they must grapple with what it means to be data-dependent in an era where privacy is paramount. Switching from user-level data collection to macro-level analysis like MMM seems like a bad choice; expect the marketing mix model to solve all budget problems. Others say it's too aggressive and not tactical enough.
But don't let the perfect be the enemy of the good. The MMM train has left the station. Marketers need to approach measurement challenges with the right mindset and approach.
