Heinen’s gets deeper insight into store operations
Chain uses machine-learning technology to flag missed sales opportunities
April 17, 2018
Cleveland-based grocer Heinen’s Inc. has enlisted a cloud-based, machine-learning solution to help identify potentially missed sales opportunities in stores.
The software, from New York-based CB4, digs into a retailer’s point-of-sales data to unearth consumer demand patterns that may indicate an in-store operational issue impacting sales of a product — at the SKU level — with high local demand. Findings are translated into actionable recommendations and sent to store managers with guidance on how to fix the issue, in turn capturing the lost sales and enhancing the customer experience.
“As a retailer, we’re continually seeking out technology to help us improve our business processes and efficiencies,” said Greg Sotka, director of category management and procurement at Heinen’s.
“We think this technology is going to help us detect unmet demand at a store level, and that could cross many different areas. It could simply be flagging out-of-stock items. It could be missing shelf tags or a sign that’s out of place,” he explained. “The system could also highlight some store-level decisions about what’s produced in the stores, such as in our foodservice areas — what they’re putting out on the shelves everyday might not always be meeting that local demand.” Product mix opportunities in certain stores, too, could be revealed through the tool, he added.
Plans call for Heinen’s store managers to train on the CB4 software in the second week of May and make it available at all 23 stores in Ohio and Illinois by the middle of the month, according to Sotka.
As a proof of concept, Heinen’s did a live demo with CB4’s solution, which includes a mobile app that provides easy access for store managers, who can also provide feedback through the tool.
“We walked some stores with [CB4], and they were pretty quickly and efficiently able to point out missed opportunities. We were surprised at the results,” Sokta said. “To be honest, we were a little skeptical at first because the patterns were not familiar to us, and they didn’t know Heinen’s at all. But we pulled up the data on their app, and 75% to 80% of the recommendations that came up were spot-on and we could see at the shelf.”
CB4’s software focuses on identifying patterns for high-demand SKUs. When a pattern is found in a store and that SKU isn’t sold at the expected levels, there’s a tangible gap in demand. Using machine learning, the solution then points to the statistically most probable operational issue to have caused that gap.
“What we do is upload the POS data from an entire chain, run that through our machine-learning algorithms and, through that, detect patterns not being met at the store,” said Matthew McAlister, director of marketing for CB4. “For example, maybe once every two weeks or so, we would run through all the data and send recommendations to the store saying, ‘These 10 items should be selling in your store in much higher volume than they actually are. So take a look at them.’ The store manager will go take a look at the items, report what was wrong, and then our algorithms learn from those responses and deliver better recommendations in the future.”
Common execution problems include promotions not properly applied, ticketing discrepancies, out-of-stocks, missing price labels, display and signage issues, damaged items or products left in a back room, among others.
“So for a grocery store like Heinen’s, we’d be delivering 10 to 15 recommendations per store every two weeks to correct in-store operational issues across their entire chain,” McAlister said.
Each recommendation delivered to a store receives a response from the store manager via the CB4 app on whether the issue was discovered or if nothing unusual was found. Return-on-investment and compliance dashboards enable managers to monitor the issues uncovered, their team’s execution status and the revenue driven.
“This is the first experience we’ve had with a machine-learning solution that takes our POS data and provides recommendations based on recognized patterns,” Heinen’s Sotka said. “It’s very easy for our store management teams to look at these recommendations quickly through the mobile app and then respond to what they’re seeing by checking the items out on the shelf. ”
CB4 reported that its clients have a 95% compliance rate on recommendations from the stores and see a net sales gain of 0.8% to 3% from its solution.
“Our store management teams haven’t worked with this yet, and we want their feedback to make sure it’s a useful tool,” Sotka said. “But the early proof-of-concept work that we did was promising.”
About the Author
You May Also Like