Root S-1 Breakdown & Commentary

Root’s going public. What follows is a quick breakdown of core metrics and a few things I find interesting. Before diving in I have to applaud the team for the truly remarkable success they’ve had to-date.

Overview:

Root is a modern, full-stack auto insurance company known for taking a telematics/UBI approach to pricing. While the average age of large auto insurance companies today is over 100 years old, Root’s built on its own 21st century infrastructure which allows it to better find, serve, and retain customers. Root believes that better pricing on a compulsory product like auto can lead to a long relationship with customers and can be a great insertion point to add more products over time.

Financials:

At the time of filing (Q2 2020), Root had 334,327 auto and 5,974 renters policies in force. Those numbers were up 51% YoY and 342% YoY, respectively. Root grew premiums 71% YoY to $608.6m, of which 47% was from renewals.

Interestingly, in this management presentation, Root expects a slow down in Q3 in both policies and premium with decreased QoQ growth. They project 328,000 policies and $597m in premium in Q3 2020 — good for 35% YoY growth in policies and 40% YoY growth in premium, respectively.

Retention:

Root’s retention rate (not including recissions) in term one is 84% and in term two is 75%. These terms are typically 6 month periods which indicates a 37% annual churn rate for Root, worse than industry averages of ~15% (although the rest of the industry isn’t growing at anywhere near the rate Root is and the majority of incumbent business is renewal).

This doesn’t tell the full story, though. The company also shares that churn increases when accounting for recissions, which would be the most apples to apples comparison with other “subscription” businesses. Recissions happen when the company rescinds business or, at times, when substandard business churns. The company indicates the increase in churn is 33% in term one and 10% in term two. If these increases are on an absolute basis, this would indicate an annual churn rate closer to 67% (33% annual retention).

Gross Margin — Insurance Margin:

Gross margin in insurance is a bit tricky to calculate. One way of doing it is to take top line premium and investment income/other revenue, subtract losses, LAE (loss adjustment expense or the cost of servicing those losses), and renewal commissions paid. This gives you a truer picture of the actual margin on the insurance business outside of capital/reinsurance costs. Looking at this number outside of COVID you see Root’s GM consistently negative — generally somewhere around -25% — although during COVID the company has been moderately positive (~3%).

Management takes a fairly similar approach to gross margin here:

Gross Margin — “Normalized”:

I’ve also seen some use a “normalized” gross margin considering the company cedes 70% of its premium to 3rd party reinsurers. This just means Root has a deal to give 70 cents of a premium dollar to a reinsurer to cover losses and minimize capital requirements. In exchange for those ceded premiums, Root receives ceding commission from its reinsurers. This appears to be on average 25% (indicated in statutory filings and management presentation).

Root also employs a captive reinsurer structure for another 15 cents of the dollar ceded, and finally a stop loss program to try to further flatten their loss needs. While it’s hard to say exactly how Root uses its captive, generally these are used to have a lower capital requirement on the holding company. A Cayman captive has a 8:1 capital requirement while Root corporate has an 3:1 capital requirement (read: Root needs to sit on less cash using a captive).

Through all of this innovative, asset-light structure, though, the company is on the hook for some amount of losses or profit incurred. If it has losses + LAE greater than 100% in the future, the business will need to use its balance sheet to pay down claims. Considering historically the combined loss ratio + LAE has been comfortably above 100%, we should further discount a normalized gross margin figure.

To get to a “normalized” gross margin number, I’m taking the average ceding commission (25%) plus/minus expected underwriting profit/loss on the 30% retained premium, as well as any impact from their quota share.

Given historical losses + LAE hovering around 120%, I’ll discount by 5% and use 20% as a gross margin number, although anywhere between 15–25% is probably reasonable with this method.

Payback:

For the sake of simplicity (and favorability to Root) we’ll use the normalized approach to calculate payback. Taking a 20% GM and the existing customer lifetime of 1.5 years dictated in the prospectus (including recissions), you can see Root has room to improve their unit economics as the implied LTV/CAC in 2019 was .41 (excluding any underwriting costs such as the cost of data or underwriting systems which is likely ~$50-$100 per app).

Simply, without meaningful cross-sell or major acquisition improvements, the business will burn significant amounts of capital for the foreseeable future.

*Normalized model (with modified assumptions) as constructed by William Blair

A few other worthy notes on the business.

• Despite the unit economics outlined above, Root is growing at a rapid rate at scale.
• Typical LTV/CAC for incumbent auto carriers is ~1.4 and closer to 2.0 for direct carriers.
• Root’s LTV is considerably lower than average given far above average churn rates either way you slice it — industry average on mature books is ~15% annually.
• Root’s $662 CAC above is elite — average incumbent is ~$950, direct auto ~$700. Note: the company reports an average (marketing only) CAC of $332, further driving this point home.

Root’s fundamental innovations:

Direct to consumer

• While this seems trivial in 2020, simply cutting out the agent in the auto insurance process is a well established playbook to drive growth. Over the past 20+ yrs, GEICO (Berkshire) and Progressive have consistently taken market share from the other agent-driven companies and are now number 2 and 3 in market share respectively.
• Root takes this direct to consumer model even further by making the purchase process faster and simpler than their incumbent competition — you don’t have to waste 15 mins trying to save 15% anymore. You can start that process from your couch.

Book-wide telematics based pricing

• Root didn’t invent telematics based pricing but it was the first to bring it to the mass market. Root allows customers to download its app, drive around for a couple of weeks to establish the type of driver they are, and then get a (often better) price on car insurance. All great consumer companies have some hook and an app that can generate a price on how you’re driving has proven to be a great hook for Root.
• Not only does this in theory attract better drivers but it has the chance to lead to lower loss ratios over time. This is incredibly desirable to Root’s reinsurance risk partners, and the capacity those risk partners have given the company is a core reason Root has been able to grow so quickly.
• Those lower loss ratios are only possible as Root will have data others may not. As Root sells a new type of insurance product and maintains a direct relationship with its customer, over time it can leverage a large, previously undeveloped data set. How its customers drive can start to lead to more intelligent pricing, and more intelligent pricing may create a hard to penetrate moat. The most attractive auto insurance customers are obviously those that don’t get in accidents, and if Root can a) target those customers more effectively and b) give those customers superior pricing, the company can sustainably generate above average returns on their investment.

Proprietary technology

• Few people truly understand the rats-nest that is the technology stack at most large insurers. This tech debt limits the speed incumbents can operate and innovate, so by building its own systems Root has a large advantage over its larger competitors.
• Even as insurance companies start to move to 3rd party software vendors they will still be rooted to their old systems and forced to operate at the pace at which they can work with a partner. Over time, this may allow Root to maintain an edge in product development and distribution.

Asset light capital structure

• While traditional insurance companies have traded off of a price-to-book multiple, Root has stripped itself from that comp set as they are much less capital intensive. This higher multiple is in large part due to lower cash requirements.
• Root leverages quota-share agreements where it locks in a gross-margin by ceding the majority of premiums to a reinsurance syndicate.

A few concerns with the model:

Unit economics

• At the end of the day, every business should be assessed on their ability to generate more than $1 on every $1 they spend. Right now, Root doesn’t. Not only is it losing money on every acquired customer, it’s losing even more money on the ongoing losses the company (and its reinsurers) will have to pay for.
• Even in the past quarter with most folks not driving due to COVID and losses being at an all time low, Root is still achieving a gross margin adjusted payback of .41 — this needs to change dramatically to grow sustainably.
• Root has also displayed a history of underpricing their products, leading to large rate increases and higher than average churn rates.
• You can see evidence of their reaction to underpricing in a fairly dramatic increase in ARPU from $805 in June 2019 to $909 in June 2020. I expect this trend to persist as the filing process takes time across states.

Competition

• Most large insurance companies have established a telematics/UBI program and will start to spend aggressively against Root. As its costs to acquire customers increase so will the pressure on unit economics. The longer Root hovers under or around a 1 LTV/CAC ratio the more its future will be tied to liberal funding and reinsurance markets.

Reinsurance pricing

• Root’s model is dependent on reinsurance capital. While it has established a captive (Root Re) to try to eliminate some wholesale transfer pricing issues, Root remains dependent on third party capital to power their model. At the relatively smaller scale it’s at today this is less of a concern, but as scale increases the concerns from reinsurers and their rates may increase as well. Should Root lose access to some of this capital it will need to decrease growth rates and take on a higher capital burden, reducing the shine they have today as an asset-light growth story.