This post is part of the Method a Month Series. Look for a new post each month and learn how our unique methods can help you with custom market research solutions and brand decision clarity.
A few weeks ago I met up with my aunt and cousin for an early dinner at a restaurant in town that is part of a national chain. We’ve eaten there before and had no problems, but this visit was an exception. I’ll admit – it was quite early so the restaurant was nearly empty and potentially understaffed. We knew what we wanted to eat and placed the order relatively quickly, but the food took forever to arrive at the table. We also had to ask for their “Signature Bread Basket” to be brought to us. When our food finally did arrive, my cousin’s order was incorrect. She wanted plain buttered noodles with sauce on the side and was served noodles in their version of Alfredo sauce. A plain bowl of noodles was quickly brought out, but they were sticking together terribly. We were given free muffins as a “we’re sorry” offering while paying the bill, and our server was very apologetic for our experience. Poor service, incorrect orders, and less than stellar food – I might be able to overlook one of those issues, but all three? How can you determine which is the most important to your customers?
Driver analysis is the answer since it can be used with a variety of market research questions, such as “What drives guest satisfaction at my restaurant?” “Which aspects of my test product most influence taste?” and “What should my customer service representatives work on to improve call quality?”
Driver analysis is used to quantify the influence an attribute has on an overall measure of performance. The Pert Group’s method for driver analysis goes beyond traditional quantitative statistical techniques, such as correlations and regressions, by utilizing an Ensemble Prediction method that incorporates Relative Weight Analysis and Machine Learning Algorithms. The drivers are reported as the confidence that the attribute is a predictor of the overall measure of performance. Scores range from 0, indicating that the attribute is not at all a driver, to 100, indicating complete confidence that an attribute is a driver. Positive values indicate the attribute predicts positive market performance and negative values indicate negative market performance. Our method also identifies the statistically unique groups of drivers in the analysis, so that you can say with certainty that one attribute is more important than another.
Do you need assistance figuring out what matters? The Pert Group can apply driver analysis across industries, in many situations, with all sorts of research questions. Contact me for more information. My aunt and I decided to chalk our poor experience up to the time we visited. She and my cousin have been back during more traditional dinner hours and their experience was again up to par.
