When “failing fast” works, and when it doesn’t: Insurtech CEO

Failing fast on small projects and establishing a feedback loop is a critical part of the innovation journey, according to one insurtech.

“You have to be able to ask your customers what’s going on – what do you like, what don’t you like?” said Jeff McCann, co-founder and CEO of Apollo Exchange, an ecommerce marketplace that enables brokers to quote, bind and issue policies entirely online. “It takes a lot of candour. We know we’re not perfect. You have to be able to ask and engage your customers, whether it’s us engaging brokers, underwriters and consumers, or whoever you consider your customer to be.”

McCann spoke last week at the Insurance-Canada.ca Executive Forum. He provided a case study of Apollo, exploring the modern insurance buying process, innovation, challenges overcome and keys to success (so far), among other topics. He stressed the importance of four factors:

• Low barrier to access (how to access insurance as easy as possible)
• Low barrier to support (getting a question answered with as little friction as possible)
• A feedback loop
• A feeling of trust and value.

For certain projects, failing fast is crucial, McCann said. “That’s the type of mentality – fail fast on short-term, small projects. It’s not fail fast on a five-million dollar implementation that didn’t work.”

As an example, McCann said Apollo once had a feature where if a broker got kicked out of the platform, the system would automatically send the application to a panel of underwriters. “We’re like, ‘This makes perfect sense. We’re doing the broker’s job for them. They are going to be so happy,’” he said. The broker would then get emails from underwriters with quotes.

“Brokers hated it,” McCann reported. “They were like, ‘We want to control the process of marketing the account and broking it out to the markets [where] we want to place it.’ We didn’t think that at all. We had no idea that would be the case.

“So, here’s something we thought was a great idea. We acted on it, we got feedback, it was a total failure and we just turned the whole thing off. We’ll iterate on it and learn from that experience.”

In general, feedback for Apollo comes in two ways: qualitatively, through a live chat feature and callings, and then through the collection of data. For example, the data can say where the drop off points are; what questions people hate; which products convert with which brokers in which regions; traffic data based on time of day, weekends, and product; drop-off points (not just quote to bind, but who started conversions and then left after seeing a price, for instance).

“If there’s a huge drop-off after that, we say, ‘Well, our price is out,’ and we go back to our underwriters and try to get it right on price,” McCann said. “Or we know if there’s a whole bunch of drop-offs, is that landing page something we can [change]?”

Being data-centric and data-driven allows that kind of feedback loop to be established and improved, McCann said. “It’s a constant evolution,” he said of the feedback loop. “How are we constantly establishing and maintaining a feedback loop that creates action items out of it?”

Why would the industry not do these things [fail fast or establish feedback loops] in the first place?

Fear of failure, suggested Donald Light, director of Celent’s North America P&C practice, in a different conference session. “We have to be willing to fail, which is… a foreign concept in virtually all existing corporations.”