Jonathan Schwartz Chief Innovation Actuary

January 2019

With the massive growth of the tech industry in Insurance ie InsurTech, we are starting to evaluate new insurance lines of business using “techy” terms and metrics. One of these is “Product Market Fit”. Is this a case where transporting a model that works in other industries won’t work in insurance?

So what exactly is PMF?

A lot has been written on the subject but I’ll quote one of my favorites – Andrew Chen.

Andrew writes that “Product/market fit is when people who know they want your product are happy with what you’re offering” he continues to write that “So what does product/market fit look like from a metrics standpoint? Really the best metric to think about is probably retention, which you’d analyze using cohorts. That means it’s sticking, and you need to figure out acquisition” because “if you have a leaky bucket, then all those users at the top of the funnel won’t help”.

A Key metric that Andrew mentions is: LTV=3*CPA

Where LTV meaning Long-Term Value, meaning all the expected profits from a customer until he churns and CPA is the Cost Per Acquisition of this customer. So basically when we have product-market fit we should be making enough from our customers to make it worthwhile to go find them in the market, otherwise, our business just won’t scale.

The usual way to product market fit is to launch a minimal viable product (MVP) and to iterate and improve it until your users can’t live without it.

Some important issues:

· What is usage in Insurance? Underwriting? Claims? What is our metric?

· We don’t always know the LTV in insurance. If we are in a new line of business it could take some time to find this out.

· If the customer “enjoys” the insurance product but it’s being sold for a premium below the claims cost — are we in product market fit? Or just subsidizing a product that can’t hold its actual cost?

· Retention. This is measured usually in years and not days or months like with mobile apps. So how long do we need to wait?

So I posted this question to the InsNerds Slack channel and to The Backstory, hosted by Andrew Chen and got some interesting responses:

· Just like any other product — conversion, CAC and time to close would be initial indicators of whether you are on a good path. Profitability is difficult with insurance and that component does take more time — but you will have a feel.

· So for new entries, you would think that they would want to go into potentially big markets. How does one compute that in insurance? Well, with losses. So the big markets would be those where there are already big (potential) losses and the new market would cover those. But the thing is, unlike any other industry, this isn’t an industry where you can just cover those losses in an insurance contract. The losses would be just too big to get enough capital to place against. SO perhaps the option is to build technology so that you can assess the risk better than current industry players.

I personally don’t really have any good answers, what do you think?