In August 2018, in our Q2 letter, I made an analogy between Ben Graham’s investment in GEICO and our transition to focusing on businesses riding technology adoption curves. I also compared GEICO’s pervasive advertising as a means of acquiring customers with a long lifetime value to software as a service (SaaS) businesses. This later became a blog post titled “GEICOs of the 21st Century.”
Lemonade is a newly public business that appears to be an actual GEICO of the 21st century. Three keys to understanding Lemonade: its customer acquisition strategy and rising lifetime value, the flywheel of machine learning, and the founders.
First, it’s helpful to remember how traditional insurance companies acquire customers: they spend millions on TV commercials in an age of declining linear TV viewership. Meanwhile, younger demographics are increasingly moving to social platforms like Snap. Many also pay commissions to agents, limiting the types of policies they can offer.
Lemonade reaches young consumers directly through digital advertising, then onboards them not through tedious agents or forms, but rather through a delightful app experience. In the process, Lemonade gathers orders of magnitude more data than traditional broker-based businesses.
One example: a standard home insurance policy is based on a form with between 20 and 50 fields (name, address, etc.). Lemonade’s onboarding bot Maya typically only asks 13 questions but, in the process, generates 1,700 data points.
This is key: that data feeds into Lemonade’s custom-built software stack (not off-the-shelf insurance-ware, or decades-old systems built on Cobol and running on IBM mainframes).
Lemonade began with renters insurance, a product with such low premiums, it’s not offered by many of its competitors. Young consumers onboard through Lemonade’s app and within seconds, they’re insured:
Given Lemonade’s lower cost of acquiring and onboarding a customer, its entry-level renters insurance is typically 50 percent cheaper than the incumbents’ offering.
Once these young customers are acquired by Lemonade, guess what happens to incumbents? They’ll never see them: about 70 percent of Lemonade’s customers are under 35 years old and 90 percent of them said they were not switching from another carrier. In Clay Christensen terms, Lemonade is competing with nonconsumption.
Over time, these young customers will graduate from renters to homeowners insurance with Lemonade and the premium will jump many-fold at zero additional CAC (this is what’s called “land and expand” in enterprise software). Meanwhile, incumbents stuck with inferior technology will turn into melting ice cubes, losing old customers and struggling to access this new digital-first demographic.
With this wealth of data and superior in-house technology, Lemonade can process claims more efficiently; today, about 30 percent of claims are completely automated. Co-founder Shai Wininger recently tweeted about how he managed to shave another second off the claims process, with many claims now handled in one second.
Lemonade’s technological lead will only widen over time: more data will result in better underwriting, which leads to lower prices; this, alongside the superior customer experience, will lower CAC and attract more customers, leading to more data, in a virtuous cycle.
At this point it should be clear that for an incumbent without a software DNA, it will be very difficult to compete: Lemonade improves its software daily. The same way that Wells Fargo will have a hard time signing up young customers to savings accounts when all those customers want is Cash App, MetLife and others won’t ever see the customers signing up for Lemonade.
The data advantage will also compound over time. A traditional insurer without extensive machine learning experience and Lemonade’s wealth of data will be hard pressed to catch up.
The final point is the quality of the founders: Dan Schreiber and Shai Wininger rank up there among the most thoughtful and smartest mission-driven founders I’ve ever encountered. They joined forces to brainstorm their next idea and once they landed on the insurance industry, they whiteboarded what the ideal customer experience should look like without any prior industry knowledge.
Initially I was very skeptical about Lemonade. How could founders with no insurance experience make a dent in such a complicated industry? But the more I investigated the origin story of Lemonade, and the more I understood how the company worked early on with industry experts to shape the ideal financial profile (capital-light, with high levels of reinsurance) and the ideal customer experience, I realized these founders are really special. And they are setting out to build a generational insurance company.
Since we began investing shortly after the IPO last July, Lemonade shares doubled, then pared back some their gains and now are up “only” 53 percent from our cost. But we expect to make many times our money over the next decade.
The company had 100k customers three years ago; today it has over 1 million. Gross profits per customer were only $0.85 back then, today they are nearly $10. Unit economics are also improving, with a ratio of lifetime value to customer acquisition cost of over 2x. This metric will expand dramatically among Lemonade’s older cohorts, as they graduate to spending more on insurance.
Lemonade’s software advantage extends to lighting up new geographies; it started business in new U.S. states (each state has its separate insurance regulator) without having any employees in those states, and the same is true about recent expansion in Europe.
The company has also begun layering on new insurance products, like pet and life insurance, and has teased the upcoming introduction of auto insurance.
Currently, customers on average spend $213 with Lemonade (for comparison, this number is $1,725 for Progressive). As the company introduces more products to cover more of its customers’ needs, we expect premium per customer to rise.
Lemonade should also be able to grow its customer base at least 10x over the next ten years. The combination of more customers, higher premiums per customer (with additional products) and a combined ratio of 83 percent (owing to better than industry-standard expense ratios) will result in an IRR from the current price of around 20 percent.
If the company expands faster—10 million customers sounds like a lot, but it’s a drop in an ocean of hundreds of millions of potential customers—this return estimate will end up being conservative.
Meanwhile, many are skeptical of the company. The stock is currently at $89 and Bank of America’s sell-side analyst has a price target of $29, while Piper Sandler thinks the company is worth $159. The short interest (number of shares sold short) is a whopping 18%.
I frequently use Twitter to test ideas and the main pushbacks I’ve seen on Lemonade are:
The valuation is too high
They are selling a commodity product, and Lemonade won’t ever beat best-in-class insurers like GEICO, Allstate, and others
From conversations I’ve had, folks complaining about the valuation are just looking at headline price to sales ratios without understanding the growth embedded in the insurance cohorts as they graduate. They also have not done the exercise of modeling the company on normalized margins over time (the company has guided to 17 percent operating margins).
While it’s true that insurance contracts are a commodity, there is nothing commoditized about how Lemonade acquires and serves its customers, or about its custom tech stack and data advantage. Rather, the company’s differentiation from incumbents is widening by the day.
These are the best contrary opinions I’ve seen, yet they are proffered by folks who have clearly spent no time actually doing the work to understand the company.
With a current market cap of $5.3 billion, Lemonade is also quite small in today’s markets; perhaps too small for most large funds to build a meaningful position.
[Note: This is an excerpt from our Q4 2020 Letter to Investors. Since I wrote this, I tweeted about the bear case on Lemonade as well as some recent disclosures we got on unit economics. I don't agree with the bear case, but it's important to understand what it is.]