When Experian’s Eric Haller put his five-slide proposal for an innovation lab in front of his boss’s boss back in 2010, the exec stopped him at the second slide and said, “I’ve wanted to do this for years, so I’ll give you the money if you can find the people.”
“I can’t remember what was on that first slide,” admits Haller, who now runs Experian’s DataLabs in San Diego, London, and San Paolo. “It probably talked about product cycles, and the competencies we needed to build it out.”
Haller’s experience managing products, strategic partnerships, and M&A for the credit bureau had made him think about the time it took to find the right acquisitions, the risk inherent in those relationships, and the overall inefficiencies inherent in the “buy” approach to innovation.
“Even though M&A will always be an important and fundamental part of our strategy, high risk, growth R&D opportunities were not always organically funded within the company.”
With that in mind, that boss’s boss (Kerry Williams, now global chief operating officer and Haller’s boss) convened the presidents of Experian’s four business units over dinner, let Haller get through all of his slides, and then hit them up for the funding. Each committed to three years, with the requirement of a monthly status meeting.
Literally the next day, Haller started working his network to recruit what he calls “navy seals of data science…the kinds of folks you find in startups,” and had his core team of 8 staff in place three weeks later. Each got an assessment of Experian’s businesses and strategic goals, and was given a week to come up with questions for meetings with the units less than a week later.
“We skipped the overviews, and fast-forwarded to Q&A and conversation,” said Haller. “It let us compile a list of opportunities for our existing data, which we distilled into a dozen or so initiatives that could get to market the fastest, and have the highest likelihood of success.”
The most promising opportunity was to pull together data for consumers who weren’t represented in its credit bureau, but could be predictive enough to make decisions. The Experian team went out to the market and got 14 deals to get data as varied as magazine subscriptions, checking accounts, and gym memberships, from which it built actionable models within a business quarter.
“Historically, we would have looked to startups to build that capability.”
Other projects followed, each illustrating a shared approach to getting solutions to the business units that can benefit customers fast. The DataLab expanded its capabilities, recruiting more experts (it now has dozens of PhDs on staff) and building its own computational resources.
“We wanted to create a sandbox for experimenting with data, but the cost was prohibitive for our business unit,” Haller said. “So we insourced the build within our DataLab and got 40x the processing power for a lower price.”
Perhaps an immediate indication of the DataLab’s success was when the business unit leaders changed those monthly status meetings to bi-annually, and then once a year. At the three year mark, the program was expanded globally.
“Our hypothesis is that we can make money versus paying six to 10 times to get returns from a startup, and that revenue is faster and more reliable,” Haller explained.
“I originally saw us as an accelerator that jumped in, got something running, and then flipped it to a business unit, but we’ve morphed into an incubator that holds onto things, and executes with operational folks, until everything is baked. Why wait for a startup to do something when we have the resources at our disposal?”