Is blitzscaling right for startup insurers?

Expect some huge growth rates in two weeks when US InsurTech carriers release their annual numbers. Root Insurance, which boasts a billion-dollar valuation, wrote $4 million of premium in 2017 and probably more than $100 million in 2018, an explosion of at least 25X. (Based on projecting forward from the first three quarters of 2018.)  Lemonade probably increased premium volume by 4X or 5X year-over-year, even though the founders say they “screwed-up an entire quarter” because their 4Q growth rate was below expectations (not a screw-up IMHO).

[Disclosure: this article contains only my personal views, not to be attributed to any company with which I am associated. Only public data is used to create this content.]

Investors are enamored with exponential growth, as evidenced by the valuations of the rapid-growth insurance companies. Investors have gotten ahead of the financials, as Matteo Carbone and I point out every quarter. The more they grow, the more the startups lose. 

Investors may also have gotten ahead of customers. Only 58% of Root customers recommend the company to a friend on Clearsurance, compared to 80% for Progressive and State Farm and, yes, 95% at Lemonade. Lemonade sports an “F” from the Better Business Bureau, apparently because they have not responded to some complaints.

LinkedIn founder Reid Hoffman and venture capitalist Chris Yeh say in a recent book that the secret to an extraordinarily valuable company is to pursue “lightning fast growth” that will “blow competitors out of the water." They call it blitzscaling, defined as follows:

“You accept the risk of making the wrong decision and willingly pay the cost of significant operating inefficiencies in exchange for the ability to move faster. These risks and costs are acceptable because the risk and cost of being too slow is even greater.”

Hoffman and Yeh cite as evidence for their concept companies such as Google, Amazon, Facebook, and Microsoft -- but not the hundreds have scaled rapidly and failed.  Failures of high-flying tech companies were once so common that there was a website dedicated to betting on which hyper-scaling startup would fail next, with more points awarded for correctly predicting bigger failures. But never mind...

Blitzscaling in insurance: don't try this at home

The word "insurance" appears in Hoffman & Yeh's book exactly once - borrowing against life insurance. As usual, insurance is different.

Blitzscaling has obvious issues for an insurance company operating in anything besides a severely dislocated market. Most fundamental is that insurance has an asymmetrical risk distribution; if the insurer is right, they keep the premium. If wrong, they lose many times the premium. Thus insurance is, by its nature, a game of avoiding “the risk of making the wrong decision.” (Venture capital is the opposite.)

More generally, as Tim O’Reilly explains, the extraordinary growth at successful blitzscaling companies was typically the result of having a great product that enables growth, not a cause of having a great product. In other words, the companies found a great product and a cash-flow-positive business model and then rapidly scaled. That enabled them to take far less venture capital money than one might think. Lemonade has already taken in 5X what Google took pre-IPO and nearly twice what Amazon raised pre-IPO.

Taking too much capital too early doesn't often end well:

The four criteria for markets where blitzscaling works. 

Criterion 1: Large market. 

In insurance? Maybe. Insurance is a $5 trillion global market, but each risk is unique, and the magic of insurance is its many niches (example). If you find yourself leading a nice niche, you may also find that cross-selling insurance is very difficult. GEICO, the #2 auto insurer, doesn’t sell its own homeowner’s policies, and Progressive (the #3) has a market share of less than 1% in homeowners. Some start-ups are trying to expand globally within a niche but are selling a product for a few dollars -- which almost never actually adds up to a large market. In other words, in insurance, the prize for being right is often not particularly large, at least not in any one year.

Criterion 2: High gross margins, so that the business will generate positive cash flow and profits when it does get to scale. 

In insurance? Rarely. Most forms of insurance, being subject to price regulation or at least intense competition, are low-margin. Home and auto insurers are lucky to make a couple of points of underwriting margin after expenses. Warren Buffet historically has treated any underwriting income as gravy. 

Some start-up insurers are saying that they will win customers today and convert them to higher-value-added products later, presumably when they are married homeowners with 2.1 kids and 2.1 cars, not broke 21 year olds buying minimal coverage only because the landlord requires it. Many years later, after one begins adulting and bundles homeowners and auto insurance, one becomes what Progressive calls a “Robinson.” Robinsons are less price sensitive, more loyal, and have lower losses - i.e. the best customers in personal lines insurance. Yet even Progressive has trouble holding onto Robinsons. As former CEO Glen Renwick explained: “For many years [we have] been what I referred to as the ‘Prep School’ for these customers. Those future Robinsons, for whom we are a leading supplier given the auto product is often the first need met, have often felt the subsequent need to shop for additional products as their lives and needs change, terminating their tenure with Progressive for reasons far from product quality or satisfaction concerns. For perspective, we have just short of 9% share of the U.S. auto insurance market, but our share of Robinsons is less than 1%.”

Criterion 3: Sustainable competitive advantage (e.g., network effects) from getting bigger faster than the competition. 

In insurance? Maybe. There may be a network effect from having better data if your product or underwriting methods are unique such that incumbents’ data is less applicable. There are scale benefits within geographic and product niches. But insurance products and methods are replicable, particularly in the US where rates, rules, and forms are publicly available. Insurance is not a two-sided market. (A two-sided market is Uber matching riders and drivers.) And insurance rarely features genuinely new business models where getting more customer feedback helps get to product-market fit faster.

There are almost no new insurance customers, so any customer has to be taken from a larger incumbent. The real value in P&C insurance is in customer loyalty, which requires a quality product and good service at a sustainable price over the long-term.

Criterion 4: Ability to bootstrap distribution.

In insurance? Rarely. Distribution for nearly all forms of insurance around the world is a challenge. See, for example, the incessant Lemonade ads in the social media feeds of anyone in New York. Most insurance execs can name the times when people were desperate for insurance - doctors buying MedMal cover in 1975, big companies buying general liability in 1986, Florida homeowners in 1993, and terrorism in late 2001. These exceptions prove the rule: insurance is sold, not bought.

However, there are a few exceptions today still, such as in commercial auto. If you’re willing to underwrite big-rig tractor-trailers or big-city for-hires, customers will find you. This is nothing new - Bermuda captives got burned in the 1970s by writing difficult-to-insure but readily-available risks such as taxis and limousines.

There are other factors that limit blitzscaling in insurance:

• Insurers consume enormous amounts of statutory capital, thus diluting founders excessively if not using retained earnings capital. An insurer’s capital is not lost but is “trapped” unless a regulator allows its release. With more MGAs converting to carriers, as I predicted two years ago, the need for statutory capital is becoming even more acute. If you've found a way around this problem, let me know.
• Regulation in insurance is stronger than almost any non-utility industry. In the US, many states require prior approval of any rate, underwriting rule, or form, and they are quick to send cease-and-desist letters (example). 
• The concept of product/market fit doesn’t work well in insurance, meaning that one could try to blitzscale straight into a business model that just doesn’t work. I’ll explore this concept in a future article.

Conclusion

Since InsurTech became a “thing” in 2015, I’ve viewed start-up insurance carriers and MGAs as a lower-risk proposition than the typical venture-backed start-up, but with lower near-term rewards too (financial speculation aside). Assuming that one wishes to build a real business and not a vehicle for financial speculation, a steady approach based on compounding over time will make founders and investors wealthy, just not tomorrow. In insurance, the conventional wisdom about growth still holds.