What is marketing mix modeling (MMM)?
Marketing mix modeling, also written marketing mix modelling, is a statistical method that estimates how much each marketing action adds to sales. It also estimates the effect of external factors, such as the season or the weather. This guide explains it to the people who decide the budget.
What it is for
- To tell base demand apart from the sales your marketing actions add.
- To compare digital media, traditional media and promotions on the same basis.
- To separate the effect of marketing from that of the season, the weather or a one-off event.
- To give marketing, sales, finance and general management the same figures.
How it works
The model relates each week's sales to that same week's marketing activity and external factors. With that relationship, it splits estimated sales into two parts: base demand and the contribution of each component.
A contribution can be negative. A strike, for example, subtracts sales. Measures such as R² show how much of the variation in sales the model explains.
What data it needs
- Sales per week and market.
- The activity of each medium, campaign and promotion.
- The external factors that affect your business, such as price, weather or season.
Two or three years of history are usual, because the model needs to observe several seasons. The data is aggregated. The method does not need your customers' personal data.
Marketing mix modeling and attribution
Digital attribution follows each user's journey to the purchase. Marketing mix modeling starts from aggregated data and covers digital and traditional media, promotions and external factors. The two methods can complement each other. You can see the full comparison.
What its limits are
- It estimates contributions. It does not prove on its own that an action caused a sale.
- It depends on the quality of the data and the variables it includes.
- A larger contribution does not mean greater efficiency. Assessing that also requires comparing activity and cost.
How to read your model with Xentinels
Xponentials builds your company's model. Xentinels lets you consult its results by market and period, and ask about them to an AI agent that answers with calculations on your data.