At Root, the price of car insurance is reported in a 6-month period, which can also be calculated from the interim reports when Policies in force is 455493 and Premiums per policy is 1616, resulting in 1.472 billion.
Policies in force naturally grows faster at Lemonade, considering that they have a wider offering of insurance products, such as home and pet insurance, from which the incoming money per policy is significantly smaller when compared to car insurance.
Root seeks only profitable growth, which is reflected in the bottom line. They deliberately terminate unprofitable customers and scale down marketing costs when it is deemed unprofitable. Lemonade, on the other hand, seeks growth regardless of the cost, and this is also visible in the increase of their marketing expenses. Lemonade has a long way to go to profitability, and much still needs to happen (many years) for it to potentially occur.
But Root would, of course, desire faster growth. Now, the independent insurance agents channel is growing promisingly, and new states are being opened, such as Washington last week. There was also a promise that their new model could increase customer lifetime value by 20 percent.
The true cost of growth efforts is revealed through LTV/CAC. LMND is transparent about those figures, while ROOT doesn’t recognize what is being talked about. There’s a good video on the topic in the LMND thread.
I must clarify that ROOT’s growth was not quite as phenomenal as the revenue would suggest. “Policies in force”, i.e., the 12-month premium income from active policies, grew “only” by approx. 20% in Q2. Approximately 8 percentage points are explained by a reinsurance agreement change, where ROOT has taken on more risk itself, thus retaining a larger portion of the premium income. For Lemonade, “policies in force” grew by approx. 30% in Q2.
I made a mistake, I should have written “premiums in force”, which grew by 20% for ROOT and 30% for Lemonade. I get some error when I try to edit the post.
Lemonade could also initially be transparent with the most important figures for insurance companies, such as the Combined Ratio, which doesn’t seem to be announced anywhere by them?
This is a bit off-topic for the thread, but I also couldn’t find the values you mentioned in Lemonade’s reports. Where could they be found officially?
Thanks for the tip. The video was incredibly long, so I won’t have time to watch it, at least not right now. Does the video concretely calculate Lemonade’s CAC and LTV? It would be interesting to see what assumptions they use. Some in the insurance industry calculate LTV by summing up insurance premiums over the entire lifetime, even though one should look at what remains below the line over the lifetime.
I also can’t find that CAC on Lemonade’s report, but does that refer to the “growth spend” divided by new contracts? Lemonade’s Q2 growth spend was 24M USD, while sales and marketing expenses in total were 59.6M USD. So 35.6M USD / 60% of the costs are related to old insurance contracts?
But I agree with Vichy that ROOT should clarify that CAC. It could also boost the stock. My guess is that ROOT’s LTV/CAC is clearly better than Lemonade’s, based on sales/marketing expenses relative to revenue and a lower loss ratio.
The video does not specifically go through LMND details. Those calculations can be found in paperbag’s videos, for example, discussed with Daniel here.
So neither company directly reports these numbers; instead, a private individual following the company calculates them, making it a somewhat misleading message that Root as a company would try to hide the figures.
Daniel’s blog is official company communication; the CEO also explains this in the first video. I would say that between these two companies, LMND has explained things more than ROOT.
One thing that makes LMND unique is this retail community, where a lot of modeling is done and things are investigated in depth.
There can be many opinions about the high price, money is certainly not free. But LMND and many investors believe this is a fair and sensible way to finance explosive growth. More specifically, LMND’s official view is in the investor day 2024 materials starting from page 163. LMND sees GC as so beneficial that a large continuation has been agreed for 2026.
ps. Now that these numbers don’t seem to be relevant for ROOT, it would be good to discuss LMND in its dedicated thread. Although undoubtedly, discussions about these competitors will continue to intersect.
Thanks for the videos - the calculation of LTV and CAC was very well covered. I took it upon myself to delve into this, and without going into Lemonade’s figures any further, I must say that “Paperbag investor” assumes a significant improvement in Lemonade’s LR in his calculations and makes significant assumptions regarding LR and retention in his LTV calculations, which have a substantial impact when calculating LTV.
I crunched ROOT’s numbers in Excel and got the following figures. I must immediately say that the accuracy of the calculations is significantly hampered by:
ROOT does not publish an exact retention rate, so it had to be calculated by comparing gross renewal premium to gross earned premium from a year ago, which means the quarterly match is not perfect, but the average gives some indication.
ROOT does not differentiate the “growth spend” portion of S&M expenses, but I assumed 70% for each quarter. This can fluctuate significantly. This also includes agents’ and similar commissions, as direct advertising costs are estimated to be very small. There can be many opinions on whether commissions should be included in customer acquisition costs.
The calculation of churn is affected by tariff increases made to insurance policies when the contract period renews. This has not been itemized, so I calculated it using “premium per policy,” which is affected by many other factors, e.g., what cars are insured, where they are located, etc.
ADR includes tariff increases, which have a retention-boosting effect. It can therefore exceed 100%. Average retention 73%.
I also calculated how many new customers came in quarterly, i.e., churn was also added to the net contract volume change. This chart shows that the inaccuracy in retention and the estimation of tariff increases yields odd figures, e.g., for the year -22.
In the LTV calculation, I used an LTM loss ratio of 65% and an average retention of 73% for each quarter. The discounter (Time value) was the same as Paperbag’s, 4%. The LTM increase in that chart thus relies only on the “premium per policy” increase.
CAC was calculated as a four-quarter average, because otherwise the fluctuations in values would be even wilder.
The LTM/CAC for the last quarters would thus settle between 3-5, which is an OK level. The loss ratio is probably no longer desired to be improved, but if retention could be improved to, for example, 80%, the LTV/CAC could rise by 1-1.5.
Is this ROOT’s solution to improving customer retention and accelerating customer acquisition? Customers coming through agents do not need to do a test drive or tie telematics to contracts. Pricing is done using traditional methods. Agents are instructed to keep quiet about telematics-tied contracts. They can only mention it if the customer specifically asks about it.
Could the idea be to convert these customers to telematics over time? A test drive can be done without the risk of a price increase (one can always stick to traditional pricing). Or do they just want to make the most of the portal they developed, which is easy for agents to use?
How do they intend to compete with this pricing model?
Presumably, the loss ratio will be higher for these customers. Will this lead to a situation where these “bad” customers, who shouldn’t take telematics in their contract, remain on traditional pricing until their payments are increased with annual raises, and then they are lost to a competitor? An interesting twist.
Isn’t calculating CAC (Customer Acquisition Cost) challenging and also more expensive in the ROOT model, because those royalties to agents run forever? According to Gemini AI overview, ROOT pays 7.5-25% commission. Apparently, those costs are in the 10Q under other insurance expense? Which is larger than marketing.
For PB, in the LMND model, the future LR (Loss Ratio) seems to be around ~70%.
That’s right, in ROOT’s case, it’s not even worth trying to estimate CAC, because we don’t even know how large a portion of current contracts came through agents. And otherwise, essential figures are undisclosed. And indeed, they are under other insurance expenses, good point. In ROOT’s case, we’ll have to settle for following the scaling of costs.
Commissions over 15% do sound quite large. That’s already the level for demanding corporate insurance policies in the US. In Finland, annual car insurance commissions average 5%. I checked a couple of US sources, and they could be 5-20% in the first year and renewal commissions 2-5%. And while commissions are certainly an expense, it must be noted that if a car insurance buyer wants to acquire insurance through an agent, the commission is included in the costs in all options presented to them. Typically, the commission is added on top of the so-called net insurance premium, and the buyer is willing to pay it in exchange for the agent’s service. But we don’t know, of course, if ROOT compensates for the commission cost in the net insurance premium to be competitive.
A small defensive victory, but a disappointment for me personally. High churn still burdens growth, although there was a small improvement compared to Q2. The CEO promised that they are striving to develop models that would allow them to abandon mandatory test drives in all channels. I, for one, am very much looking forward to this. This has now been abandoned in the agent channel, at least. I don’t understand why it can’t be done in all channels already. Do agents get better information about the customer and claims history, etc.?
On the positive side:
3x growth in the agent channel
Gross written premium growth Q/Q 12%, whereas in Q2 there was a 14% decrease. The timing of insurance periods may, of course, also affect this.
Noted acceleration of growth in October and promises of better overall growth in Q4
Only 10% of US agents distribute ROOT’s insurance, meaning there is plenty of room for growth.
