On Wednesday, April 17, the Chicago Connectory hosted Alexandra Gaski, a data scientist at McDonald’s. This event focused on how companies can use data to determine the success of their marketing and media efforts.
Gaski began the presentation by identifying issues all companies face when it comes to their marketing and media. Leadership will typically ask how marketing dollars are contributing to the overall business or how marketing directly influences sales. However, traditional marketing metrics, like in-store sales or media metrics, do not answer these questions.
Gaski explained that since there are so many external factors influencing sales and purchases, it is extremely difficult to attribute a sale to a specific marketing campaign. This is especially difficult for McDonald’s since the majority of their sales are in-store purchases compared to online sales.
However, by identifying all of these potential external factors (weather, competitor ads, region, size of store location, etc.) and using a complex regression model called Marketing Mix Modeling (MMM), McDonald’s is now able to determine the impact on sales and ROI measures for individual marketing activities. This model can be used not only to understand past performance but also to optimize future media spend across channels, products and messages.
While these results sound wonderful, it was no easy task. Developing these types of insights takes a great deal of planning with various internal media teams. Everyone within the organization needed to coordinate to insure proper data collection.
During the last few statements, Gaski provided insights on how organizations can start planning for larger data science projects. First, start gathering all of the data in your data “world”. This includes external economic variables, company variables and marketing variables. Second, hire a data scientist or analytics firm – these projects are not something you want to do on your own. Third, create a culture around measurement. Lastly, benchmark against yourself. If you are McDonald’s, it does not make sense to compare your results to Amazon, for example.