Insurance technology expert Lemonade (LMND 1.15%) was a high-flying market darling once upon a time. Its fully automated end-to-end process, with artificial intelligence (AI) tools running everything from applications to loss payouts, was a perfect match for the increasingly automated trends of the early pandemic era.
But the business grew too fast for its own good. Lemonade's financial statements turned sour, and the stock took a swan dive. The market cap reached an all-time high of $10.4 billion in January 2021. The company is worth only $950 million today.
Is the company poised for a comeback, or are the lemons really rotting? Let's take a look.
Lemonade's unique plan: Start from AI tech to build an insurance company
Lemonade's long-term prospects depend on one ambitious goal. This technology company wants to become a robust insurance service before today's burly insurance companies can copy Lemonade's efficient technology.
I'm not making this up. Lemonade CEO Daniel Schreiber hammers this point home as often as he can. In the earnings call for Lemonade's fourth quarter of 2021, for instance, Schreiber stated his case this way.
"Our long-standing, two-pronged strategy has been to win with technology and to grow with our customers," he said. "That is, to build a digitally native company on the premise that an insurance company built on a technological foundation will be able to serve its customers and quantify risk with a degree of precision, another level of automation unavailable to incumbents."
Ten months later, at the company's annual investor day in November, Schreiber explained why the incumbent insurance leaders are ill equipped to do what Lemonade does.
First, he presented the analytic capability of the unheard-of GPT-3 system from OpenAI (two weeks before the GPT-3-based ChatGPT went viral). Schreiber explained how GPT-3 had learned how to write sonnets about insurance by making trillions of connections between data points in a massive database -- and how Lemonade's deep learning systems are doing the same thing in a large and ever-growing database of insurance data.
"What we are showing you today is different not in degree but in kind. The kind of connections, the kind of systems upon which we are built, nobody else can do. Unless you will build this way from the ground up, you just can't make the connections, and it's through no fault of anybody else," Schreiber said. "If you founded your company in the era of the horse-drawn carriage, you optimize for that era, and you find yourself flat-footed going into the digital era. We have the good fortune of being founded in the digital era, and therefore, we built it the way we built it."
In other words, Lemonade's leaders believe that the company has a unique and hard-to-copy advantage over traditional insurance carriers. The system is far from perfect, but it should outperform the incumbents on every insurance metric that matters in the long run. The AI-powered lemon grove just needs to put down deeper roots into a rapidly growing collection of insurance events and customer data. That's the nourishment Lemonade's AI-driven business runs on.
No short-term growing pain, no AI-building gain
Lemonade's renters and homeowners insurance plans have always been affordable and easy to use, as long as you don't mind managing your insurance coverage online. As a result, people signed up in droves before Lemonade's machine learning systems had enough training data. As a result, the company may have approved too many insurance applications and set its premium fees too low. The company was also unprepared for an abnormal number of natural disasters over the last two years. A more mature AI platform, evolved from a gentler rate of client growth, would have been prepared to make better decisions.
That's why Lemonade's loss ratio is so high, which in turn explains why the stock price has fallen so fast. Lower is better for this metric, which measures the proportion of insurance premiums that are paid out to cover insurance claims. Lemonade aims for a ratio of 75% or less in the long run. In the second quarter, that ratio spiked to a harsh 96%. This figure shows the company's struggle to balance premiums with payouts, an essential aspect of a profitable insurance business.
On the upside, 1.9 million customers are building Lemonade's fundamental database of insurance events very quickly. If the company's basic business model makes any sense at all, the benefits of higher-quality training should start making Lemonade's insurance process respectably effective soon enough.
More sugar is on the way, as the business is heading in a more profitable direction. Lemonade's loss ratio is consistently higher in its newest service lines, such as car insurance, and below 50% in the more mature renters insurance segment. It seems that more data results in more effective insurance decisions by the AI systems.
And beyond that, I would look to the world of competitive chess as a model for the insurance industry's future. In the mid-1990s, human grandmasters wiped the floor with the best chess computers. Nowadays, chess computers are essentially invincible. The chess app on your phone would crush world champions Magnus Carlsen or Garry Kasparov without breaking a digital sweat.
Can Lemonade checkmate the massive insurance industry?
So here's how I see it. Investing in Lemonade today is like betting on chess computers in the 1990s. It's not a guaranteed win -- no investment ever is -- but the company can lean on a debt-free balance sheet with $942 million of cash equivalents and short-term investments while the AI systems learn how to run a profitable insurance business.
The long-term opportunity is like turning a pawn into a queen; even a 1% slice of the multitrillion-dollar insurance market would be a king-size boost from $257 million of annual sales today.
In other words, Lemonade is a promising buy at these low prices, as long as you expect the good times to start rolling within the next two or three years.