Most shoes that Americans wear are made in Asia and have a six-month cycle from design to sale. That process, including the manufacture of the shoes, has been the same for 60 years, according to David Roberts, VP of enterprise transformation at Under Armour. The retailer, which gained fame for disrupting workout clothing, is trying to do the same with shoe sales.

Under Armour is now selling shoes with 3-D printed midsoles directly to consumers—shoes that are made on demand after an online order and shipped within two days. That’s a culmination of a transformation process that began with Under Armour purchasing many of the major players in the fitness tracking app market, then connecting those apps to the shoes it was selling to create a treasure trove of Internet of Things (IoT) data. The result? A shoe built with the collected knowledge gained from previous shoes.

“We wouldn’t be able to do this without the data and the insights that we glean from [fitness apps],” said Roberts, who spoke at the 2018 BI + Analytics conference in Huntington Beach, Calif. He adds that one example of an insight Under Armour gathered from the connected shoes is that its customers often use cross-training shoes for running.

Roberts explained that the 3-D printed shoes can be produced at 2,400 per hour, compared with 2,400 per eight hours at factories. The cost to Under Armour moves from $24 per pair out of the factory to 30 cents per pair using the on-demand 3-D printing method. Interestingly, the savings doesn’t seem to have made it to consumers just yet, as the shoes with 3-D printed midsoles retail for almost $300, well above the average factory-made Under Armour shoe. Perhaps the initial cost of the 3-D printers affect that price in some way.

Analytics Drive Transformation

In using analytics to build products, Under Armour analyzes its own data from its SAP system but adds non-SAP sources to get a more thorough picture of consumers.

“Much of our operations runs on SAP,” said Roberts, “But more important is that much of the data is outside of SAP and outside of Under Armour.”

Under Armour uses SAP Vora to connect its Hadoop cluster layer to its SAP HANA instance. On top of that, the retailer leverages SAP Digital Boardroom, SAP BusinessObjects Design Studio, and Crystal Reports for analysis.

“You want to make sure to give the business the tools to answer questions they didn’t think to ask—and make sure they are not beholden to the IT team to determine what questions should be answered,” asserted Roberts. “That’s changed the business, it’s revolutionized what’s been done.”

While building the design for its 3-D printed shoe, Under Armour changed from its traditional approach of seeking feedback directly from consumers. Instead, it took to social media analysis using the natural language processing capabilities of SAP HANA. Essentially, Roberts said, the final product was combining all the successful pieces of previous Under Armour shoe designs.

People Still Matter in Analytics

Agile development drives Under Armour’s transformation—the company works in six-week investment cycles. The operating model that drove the fitness app acquisitions was built in six weeks, and now produces five percent of the company’s total revenue.

Under Armour also scores the data it uses by source—not all data is created equal. For example, the highest level of trust goes to SAP ERP data followed by data that the retailer pays to acquire. On the other end is social media data—which isn’t worthless, but Under Armour tries to take the necessary grain of salt when using it for analysis that drives product decisions. That’s where the people, not just the machines, are still important.

“At the end of the day, you still have to have a person look at the analytics,” said Roberts. “Technology gets you to the table, but you have to decide how to break the data down.”