How to determine which stores benefit from media spend — and which do not
Successfully identifying which stores and products do and do not respond to media stimuli can help you understand the “true” drivers of your business. In this case, we used Media Impulse Shock Analysis to identify which of our client’s stores were positively impacted by media stimuli, by channel and to what extent.
The problem: Our client was running media in certain parts of the country and wasn’t seeing any impact on the business, which was unexpected. When we looked at the data in aggregate, we realized that the positive performance of some individual stores was being washed out by stores that were doing poorly. The net impact of media stimuli on the business was inconclusive.
We needed to dive deeper into the data to find out what was going on at the individual store level. A traditional media attribution model would not have been sufficient. In fact, our previous blog posts and podcasts explain our rationale for WHY WE BELIEVE ATTRIBUTION MODELS ARE INFERIOR and should be avoided.
The solution: As part of the client’s annual media planning process, we unleashed our data analytics team armed with the Media Impulse Shock Analysis (example at left) to identify which QSR stores were positively impacted by media stimuli, by channel and to what extent.
Once these stores were successfully identified, we were able to start the process of storytelling and filling in the missing pieces. We drilled down on commonalities and differences through layering in geographic and economic factors, and consumer demographics and profiles.
Throughout this process, we leveraged our INTELLIGENCE PLATFORM, statistical modeling, threshold pressure testing, and different mathematical techniques to identify the stores—with a high degree of confidence—that were impacted by media, negatively or positively. This is how we discovered that a large portion of their marketing spend was being wasted on store locations that were receiving no benefit from it.
Depending on how you looked at the data, anywhere from 35-50% of stores realized no benefit from any of the media that had been run throughout roughly the past two years.
Of course, we also identified which stores were being impacted by media. This made it easier to look for patterns in the data — characteristics these stores had in common — so that we could build a profile of the benefiting store locations.
Once the benefiting store characteristics had been identified, the client was able to reevaluate their marketing approach based on the characteristics they could control, like pricing structure and operations.