HOME | NEWSROOM | PRESS RELEASES

MIT Sloan professor uses new method to identify consumer demand trends and improve profits

CAMBRIDGE, Mass., Oct. 13, 2017 – In marketing, personalization is a big trend. But can demand be predicted at a more personalized level? Without social media data, can customer-to-customer trends be identified to optimize personalized promotions? A recent study by MIT Sloan School of Management Prof. Georgia Perakis found not only that these trends can be predicted, but that they also can be used to optimize targeted promotions and improve profits by an average of 9.4%.

“If we have access to social media data, then we can detect connections and identify influencers. However, retailers often don’t have access to the right level of detailed data needed without paying large fees, so an important question is how they can detect trends using more traditional data. Further, how can they use this information to determine the best personalized promotion strategies?” explains Perakis.

Tackling these questions in a study with MIT Operations Research Center PhD students Lennart Baardman and Tamar Cohen and collaborators from Oracle Retail, the first step was to create a customer demand model that incorporates customer-to-customer trends and influences. Applying that model, Perakis and her colleagues used the information about customer demand to make decisions about promotions, which led to increased profits between 5-12%.

The model analyzed traditional factors like store locations, the types of people who shopped at these locations, the timing of when a specific product was purchased, and how much each customer spent. “We looked at consumers who are similar and then looked at different individuals to extract groups and identify frequent patterns of buying behavior,” says Perakis, noting that they applied this analysis to males and females, different geographies, and multiple products.

They found that customer-to-customer trends can be dependent on item, time, and promotion. They also found interesting patterns based on geography. Analyzing stores of a large retailer in Ohio, the study looked at two seemingly distinct large areas: Cleveland and Dayton South. While an obvious promotion strategy would be to target the bigger city first (Cleveland), the researchers found that a third area, Columbus, was an influential hub for Dayton South, which then influenced Cleveland.

“We realized that it would be better to target Columbus first, rather than Cleveland or Dayton South, because it was like hitting two birds with one stone,” observes Perakis. “In other words, if you can motivate a group of key influencers to buy more, they can get other people to buy the same products so you can get more bang for your buck.”

She adds, “It’s clear that even without social media data, you can still understand the influence of certain individuals and groups and use that information to target and motivate others to improve overall profitability.”

The MIT Sloan School of Management is where smart, independent leaders come together to solve problems, create new organizations, and improve the world. Learn more at mitsloan.mit.edu.