Marketing Mix Modeling: ROI Fortune Teller?
November 18, 2009
In these challenging times, when every marketing expenditure must be justified, more advertisers are turning to Marketing Mix Modeling (MMM) to determine the effectiveness of their marketing plans. Empower MediaMarketing has worked with a number of clients and modeling companies over the years and helps clients through these projects.
MMM uses advanced statistical techniques to compare marketing expenditures with sales results to find causality and determine payout. It then builds a model that will predict how future marketing expenditures will impact sales. This form of analytics takes marketing metrics to the next level, enabling better forecasting of results in various marketing scenarios. Since determining ROI for past marketing efforts is only one part of the picture, MMM helps answer the question, “What will work in the future?”
MMM Pitfalls and Possibilities
Marketing professionals need to be aware of some of the pitfalls that can occur with MMM. Consider this top line list when embarking on a predictive modeling project:
-
Two Years of Data: Your modeling company should ask for at least two years of data. They should allow for the lagging impact of advertising, autocorrelation between variables and external factors like weather.
-
Clean, Granular Data: Make sure your data is clean and granular. This includes sales data and key performance indicators by week and by market (or store), as well as detailed media and marketing data. As they say . . . . “Garbage in, garbage out.”
-
Creative Details: If you run several different pieces of creative, be sure to have the copy tested and the persuasion scores included in the model.
-
Online and Offline: Digital media metrics can be an important part of the model. Make sure the appropriate measurements are included.
-
Not a Crystal Ball: Remember that most MMMs are predictive of future trends, based on the past. It may be difficult or impossible to predict results when introducing a new media vehicle to the mix.
-
Calculated Risk: Commit to using the model, but not to assuming 100 percent accuracy. Even the best model won’t be perfectly predictive. Don’t let a focus on analytical results keep you from using intuition and creativity.
While Empower is not a Marketing Mix Modeling company, we have supplied clients’ granular media data and worked with numerous analytical partners for many years. This allows us to help clients with their MMM needs by asking the right questions, assisting with the choice of a modeling partner and challenging the inputs to improve the accuracy of the output. Our research strengths allow us to lead marketing research and track social media conversations to clarify consumer attitudes and beliefs, and Empower can test your creative in concept, storyboard, or finished form. Finally, we can help advertisers determine how to best use the learning that is derived from their models.
By: Julie Pahutski, Senior Vice President of Consumer Insights