ROOT - Root Insurance

[Broeb22]

Does anyone have any direct experience with Root insurance?

 

I tried to sign up a year ago but it doesn’t exist where I live. I understand the telematics piece of it where they leverage the user’s cell phone to determine how risky their driving is and charge them based on their risk profile. I believe Progressive offers something like this (with an attachment that goes on the A/C vent or something) although I don’t think it was as widely adopted.

 

Based on the way Root describes their offering they believe low risk drivers are overcharged based on their individual risk profiles, so it seems like they may be targeting that segment of drivers. Does this have any potential impacts on Progressive, Geico, and other auto insurers?

 

Finally at a $6B valuation and 6x book value, investors are really giving the, the benefit of the doubt. Does anyone have any way to think about how this business could scale?

[gg]

How they can scale??

 

The old "lose money on every sale" and then make it up on volume. Their metrics are horrific, but the tech koolaid spigots contonue to flow and investors will eventually be surprised to find out that writing insurance at a negative gross margin is not difficult to do!

[Broeb22]

I’m trying to be open-minded about what could happen with this business, but based on your response you probably slavishly cling to buying dogshit companies and still can’t figure out why your investing strategy is underperforming for the last 20-some years. It’s also worth pointing out that GEICO itself got in trouble for underpricing insurance in the 70’s. Does that make its business model any less powerful? Could Root get themselves into trouble despite a (potentially) fundamentally better business model? That said, Root does appear to be underpricing their policies at present.

 

I’m trying to understand whether there are pieces in place here that have staying power, and absent a really profitable model, will they be a pesky thorn in the side to Progressive/GEICO? Are their non-loss and LAE expenses such that they can more efficiently price policies than others (perhaps even GEICO/PGR) over time.

 

They discuss the amount spent (wasted?) on customer acquisition in insurance. They seem to imply they can acquire customers more cheaply than GEICO/Progressive. Is there anything to that claim? Seems that would be difficult to do long-term, though while they’re small maybe that’s not a big concern. They are spending a lot on marketing as a % of sales right now though. I think I found about it because they were all over social media and I was curious what it was.

[KJP]

I have no experience with Root, but there's a discussion of it here that my interest you:  https://netinterest.substack.com/p/the-end-of-insurance

[lnofeisone]

I have no experience with ROOT. I did however dig into them for a bit. Their reviews are absolutely terrible on the customer acquisition side. The user reported experienced with ROOT is 50% positive, compared to 90% positive for Lemonade (disclosure, I'm short Lemonade). It looks like they lure you in with a lower rate and then increase the rate. Basically, seems like bait and switch. Of course, they won't tell you why a model increased someone's rate but since they have 200+ additional data points, they can blame it on any of them.

 

Their claims process seems to be easy and well managed. People tend to write good reviews on it.

 

From a data science perspective, I am very skeptical that the addition of 200 more points will somehow get them to causality (which is a holy grail for most data scientists). I do think they will collect all this data and will realize that there are probably 10-15 variables that can really define a customer. With 200 extra variables, they will have a more accurate view of a customer but is it useful?

 

I am not convinced their strategy is a big differentiator. Looking for excellent customers and offering them low rates is basically a dream for every insurance company. Geicos of the world have the scale and the ability to absorb freak events. If they are as good as they say they are in data science, they should look for customers that Geico or Progressive would deem high risk but in reality, these customers are low risk (basically Capital One model).

[gg]

I’m trying to be open-minded about what could happen with this business, but based on your response you probably slavishly cling to buying dogshit companies and still can’t figure out why your investing strategy is underperforming for the last 20-some years.

 

This is a pretty odd comment on many levels -- but let's just say that I'm of the view that if you choose to rely on a negative personal attack in a debate on the merits of a financial investment, then it says more about the lack of noteworthy reasons to purchase the stock than it does about anything else. Regardless, I'm happy to engage on Root.

 

Car insurance companies like Root are pretty simple. They're a combination of 3 lines of business that put capital at risk: indemnity, marketing, and investing. Indemnity is the underwriting risks that they choose to take. Are they selling insurance policies at a price-point where they will not pay out more in claims than what they've billed and collected their customers for? Marketing is like any other business, "can they acquire and retain their customers at a cost basis that is competitive and that leaves them with a meaningful margin after the direct expenses of the unit economics?" Investing - will the aggregated funds from prepaid policies -- or float -- enable the insurance company's equity owners to accrue a meaningful financial benefit via this interest-free loan?

 

The primary bull case that gets used to advocate for Root is that their technology allows them to compete on indemnity. In Root's case that would be the suggestion that:

(1) they can price driver risk more accurately (meaning they'll ensure that they're customers pay enough to more than cover the cost to insure them), and

(2) that they can use technology and automation to process claims efficiently to drive down operational expenses and be able to offer their product more cheaply in scale than competitors.

 

I don't think any reasonable person would suggest that Root has a moat - nor a potential moat - when it comes to the investing portion of running an insurance company. And when it comes to the marketing, I am not under the impression that they are doing anything particularly unique there -- the competition for share of voice amongst car insurance companies is about as high as brand competition gets (see any football game's commercials, or google car insurance to understand how deeply competitive that world is. Not just among the brands themselves, but also throw in lead aggregators!).

 

So the question boils down to: do they have a better solution than other insurance companies in terms of cracking the "indemnity code"? And if they do have the better pricing model than all of their competitors, does that improvement enable them to carve a meaningful and profitable position in the industry? (maybe the savings via improved indemnity are so large that they can put up a formidable fight against their competition by being able to spend more than competition on marketing?)

 

So without getting into that discussion too much, I would wonder if the state departments of insurance that regulate insurance pricing allow them to use this alleged phenomenal pricing model (worthwhile review on some of this - https://www.casact.org/pubs/white-papers/Insurance_rating_variables_white_paper.pdf)? Is the pricing model proprietary? Typically, the variables must be filed publicly with insurance departments, for all competitors to see. Is the total difference between a good pricing model and a great pricing model enough of the cost-structure that it makes all the difference? If all of the benefit is related to telematics, do they lose money on the first contract (priced before telematics has data on the driver), and then make up for it all on the renewal?

 

Ultimately, my fundamental view on insurance is that pricing models are not truly part of a car insurance company's proprietary and competitive moat. Ajit Jain can have an edge in being willing to take indemnity risk with things that other won't, and that, among other things, allows him to charge a real premium versus the risks he's taking. But with car insurance, it's much more of a commodity - the agreements are short term so the financial viability of your counter-party isn't too important / switching costs are low and you can change your policy to a different insurer in 15 minutes or less! / most shoppers simply try to minimize price versus any qualitative distinction.

 

What I think has happened here is that many venture capitalists and retail investors that have seen tech used to disrupt other industries think that that's what is going on to the car insurance industry from Root. I think Root has seized on that opportunity to sell a story of having found a better mousetrap. So VCs (pre-IPO) and public investors (post-IPO) are subsidizing the sale of a commodity product, and all this is what's playing out with the insurance losses and the larger and larger losses as they've grown. Meanwhile, the investors are excited about growth! But again, they're selling a commodity -- if you're willing to sell a commodity for a loss, your sales will obviously grow.

[Broeb22]

Gg, I don’t have a view on the stock which is why I wrote the post in the first place.

 

Thank you to KJP and inofeisone for their contributions and links.