A game of inches

Long before Openly began selling home insurance, co-founders Ty Harris and Matt Wielbut kept getting the same feedback from potential investors: “Can you distill your idea to one succinct, sexy thing?” They couldn’t. In the fintech space, insurance is as old as the Babylonians. “We hadn’t invented anything completely new,” recalls Harris. “We hadn’t patented some massive piece of technology.” What they had, and what admittedly took time to explain, was a plan to make home insurance better “in a thousand incremental ways”—and then to keep making it better. “That was our one tool against the incumbents: the ability to constantly improve,” Harris says. “And we knew that if we kept improving, we’d win.”

Six years later, Openly has raised over $250 million and sells insurance in twenty-three states. Its product is built on two core tenets: customers want better coverage with less fine print, and independent insurance agents are an underappreciated channel to reach those customers. While most of the industry has shifted direct-to-consumer, burning millions on advertising, Harris and Wielbut have taken the contrarian view that knowledgeable agents are still the best way to sell a complicated product—if they have radically-improved tools.?

Given regulatory constraints and the financial demands of underwriting, they knew it would be a long path to actually selling insurance. But that freed them to start with a blank sheet and design both customer-facing and back-end systems for continual innovation. After more than a decade of experience in the upper ranks of an insurance behemoth (Liberty Mutual in Harris’s case), and a financial giant (Goldman Sachs in Wielbut’s), they knew how slow those large players could be to change. Harris and Wielbut took that as their tactical challenge. Home insurance had boundless room for refinement—and that work would never be truly finished.?

“I fell in love with the greenfield-ness of it,” recalls Wielbut. He had worked as a programmer at Goldman, where in-house tech was a powerful advantage. He then founded a small insurance agency with Yana Harris, an MIT grad, actuary, and his co-founder Harris’s wife. They saw firsthand how effective agents were at educating their customers and how little the old guard had done to make their lives easier. Underwriting a new policy requires answering hundreds of questions—from the roof material to the type of wiring in the walls. With ten agents coding new policies every day, Wielbut and Yana Harris had a front-row seat to their frustrations. “I could just stand next to them and see a hundred clicks—the same clicks every time,” says Wielbut. “And we're like, why is this happening? We can automate all this!”??

Harris, for his part, knew that no one else in the industry was tackling this kind of holistic innovation. In his prior role at the large insurer that employed about 50,000 people globally, Harris oversaw a team of 800 people that was responsible for pricing and underwriting 10 million customers and $20 billion in premiums. Across the insurance industry, almost 40% of revenue goes to administrative expenses rather than claims or profit. Often lacking cohesive technology, the insurance giants mainly depend on a slew of vendors and perennial tangles as they bolt their tools and platforms together. For Openly, making a better UX for agents was part of the challenge—but so was re-imagining the underwriting models that properly price every policy.?

“We didn’t go into this thinking we were going to have to build it all, at least in the beginning,” says Wielbut. “We had that MVP mentality. But the deeper we got, we realized that there was nothing off the shelf that we could really use.” Having seen their old employers bogged down under the weight of their aging systems, Openly’s goal was flexibility. Without the archive of data the legacy companies use to price risk, Harris and Wielbut had to define a general language that would allow their new models to be rapidly tweaked over time. “Above the waterline, it's a beautiful, simple, fast experience for the agent and customers,” Harris says. ”Below the waterline, it's seven or eight big models making decisions really quickly, in an interactive way, that humans often make at other companies.”?

Jason Bucholz, Openly’s VP of Insurance Product Management, can see the difference. Without the constraints of old corporate structures, Openly built constant innovation into its organizational design. Insurance product managers—the people with a deep knowledge of this highly-regulated product—sit with the data scientists, who seek out new information sources, and the engineers who will actually be programming the models and their interconnections. The size of a property, for example, might be inherently linked to the potential for wind damage. But the more finely grained that relationship can be understood, the more accurately it can be reflected not only in the underwriting but also in the rating—offering the appropriate price to the appropriate risk.

In a traditional organization, each new variable would be incorporated into the model incrementally like a waterfall—moving in one direction from pricing to product to engineering. “At Openly, it all happens simultaneously,” says Bucholz. ”We get everyone together, and we say, ‘Hey, these are the variables, the characteristics that we're going to look at—and everyone's together having that discussion.” That iteration happens fast. Openly proactively makes continuous small updates to its programs and underlying models and targets a major upgrade every six months, dependent on state requirements and within regulatory frameworks. This is 10x the pace of the existing giants, who might typically upgrade their major models every five to seven years.?

The online portals used by both agents and customers get a similar treatment. Danielle Wyman, VP of Product and Strategy, works alongside UX designers and researchers to refine each step of the workflow. ”Our value proposition is, it's easy for the agents,” Wyman says. “They can get a quote in fifteen seconds. So they are not spending time re-entering data—they’re spending time talking with their customers and understanding their needs.” To make sure that’s being done right, they interview agents about how they work, track their cursor movements, and organize broader “agency councils” for discussion. The tweaking never ends.?

By Harris and Wielbut’s founding logic, there’s no limit to improvements. The core product—not just the insurance, but the tools used to sell it—will never be done. Along with typical performance indicators, Openly’s teams are organized around actions rather than functions—verbs rather than nouns. Like, “Make life easy for agents.” “Lower expense.”? “Match risk to rate.”?

“We decided early on that our pillars should be delta—things we can change,” says Wielbut, “as opposed to just a static noun.” As old as insurance is, the attitude is that there’s room for refinement—of every piece of the business. And then, every quarter, they reevaluate those pillars, changing the axis of change itself.?