Marketers need some reassuring economic news that consumers will continue to spend, even after the holidays. A new way to use data and modeling that predicts advertising conversions might bring it all into focus.
Michael Lee, senior analyst of data science and analytics at Dstillery, led a project that began in March 2018, analyzing the nuances of where and why Gap might close its retail stores.
Gap has been struggling. In the most recent quarter, same-store sales for the Gap brand fell 7%. Sales at its sister companies -- Old Navy and Banana Republic -- rose 4% and 2%, respectively, according to company data.
Using data from search, proprietary programmatic, and third-party longitude and latitude, the team chose Gap based on the number of store closures across the United States in economically healthy and ailing regions, and the fact that the retailer has never been a Dstillery client.
The idea was to analyze the relationship between consumers and stores directly in areas with physical locations that remain open and closed to attain a better understanding of the differences. To do that, Dstillery analysts built a model that is typically used to predict advertising conversions.
The analysts began looking at the income of the surrounding areas, but quickly realized that analyzing internet browsing and search data was a much more accurate predictor of whether a specific Gap store would close or remain open.
The plan was to achieve a better understanding of the web-browsing patterns of people who live close to the stores that closed. It turns out that people in those areas have a greater affinity for stores like Carhartt, LL Bean, and The North Face. They also tend to visit more traditional fashion outlets such as Chicos, Dress Barn, and Talbots, and shop more frequently at youth-style retailers like Aeropostale, Journeys, and Zumiez.
The analysis suggests companies need to pay more attention to brand data other than store and online purchases to get a better idea of consumer behavior, Lee said.
Gilad Barash, director of analytics at Dstillery, suggests measuring other types of online activity other than the physical locations they visit. “It’s a blind spot,” he said. “Other types of data can help inform strategies, ranging from where stores should open to different ways to engage with potential consumers.
There are challenges to this method. Lack of visitation data to the physical stores is the biggest limitation, Barash said. It’s also not clear whether store closures were a result of consumers in the area simply preferring to buy online, rather than going into a physical store.
Lee and Barash stopped short of predicting the exact location of future Gap closures based on challenges. The two pointed to missing data such as foot traffic and specific economic conditions, but did agree that if they had the additional data from a brand they could predict the future of physical locations.
The duo did admit that the analysis identified certain areas across the U.S. where consumers demonstrate the same type of behavior where previous Gap stores closed. Alabama, North Florida, Colorado, and Central Connecticut were some of the locations.