Shorts were active yesterday, continuing their shorting.
https://x.com/bradsferguson/status/1986123157795659882?s=46
LMND has announced that it will use, if I recall correctly, $200M in synthetic agents funding in 2026, meaning there will be no slowing down in terms of customer acquisition and growth.
Vuosi sitten kirjoitin tÀnne viimeksi ja voi veljet mikÀ vuosi tÀmÀ onkaan ollut.
Muistutan tÀhÀn alkuun vielÀ:
Yhtiö on ohjeistanut jo vuodesta 2022 asti, ettÀ he ovat adj ebitda -positiivisia vuoden 2026 loppuun mennessÀ. TÀmÀn toistivat jÀlleen eilen: Q4 2026 on adj ebitda pos.
Ovat myös ohjeistaneet yhtÀ pitkÀÀn, ettÀ saavuttavat ensimmÀisen nettotuloksen vuoden 2027 aikana. Arvioin itse, ettÀ tÀmÀ tapahtuu joko Q3 / Q4 2027.
Siihen asti book value laskee.
Siihen asti CR on yli 100%. Kunnes se ei enÀÀ olekaan.
Vietin viimeisen tunnin lukiessani tÀmÀn ketjun alusta loppuun. Katsotaanpa miten asiat ovat kehittyneet vuoden aikana.
Olet. On GAAPin mukaista, ettĂ€ SBC lisĂ€tÀÀn takaisin operatiiviseen kassavirtaan, koska kyseessĂ€ on non-cash expense. LisĂ€ksi Lemonade on ollut GAAP-kassavirta+ Q3â24, Q4â24, Q2â25, Q3â25.
Lemonade pienensi QS:ÀÀ jĂ€lleenvakuuttajille 1.7.2025 alkaen 55%â>20%. Edelleen on pystyssĂ€. Kuinka pian uskot konkurssin koittavan?
NÀkyykö tÀmÀ liian matala hinnoittelu oheisessa kuvassa?
Ja niitÀ kustannussÀÀstöjÀ aletaan nyt siirtÀmÀÀn asiakkaille.
âStarting to capitalize on our AI UW/pricing advantage. With Loss Ratio hitting our lowest levels ever, we can now turn it into a lever rather than a target, lowering costs and achieving price leadership.â (Shai Wininger, Q3 2025 twiitti)
TĂ€mĂ€kin on tehty onnistuneesti. â..And finally, all else equal, less quota share increases regulatory capital needs. However, with an improved loss ratio and the expanded use of our wholly owned captive, we are able to offset these pressures such that there is no material change in our capital planning.â (CFO Tim Bixby, Q2 2025 earnings call)
Mielenkiintoista. TTM gross loss ratio on nyt 67% ja Q3 net loss ratio oli 64%. TÀmÀ tapahtui 9kk postauksesi jÀlkeen.
Voin jo nyt paljastaa, ettĂ€ Lemonaden liikevaihto (ei siis âliikevaihtoâ) minimissÀÀn triplaantuu 2027 vs. 2024. 3x kolmessa vuodessa.
Jep, keksivÀt, ettÀ loss ratiot laskevat ja esim. LAE (7%) on jo parempi kuin vanhoilla toimijoilla. GC tosiaan halusi jatkaa ja laajentaa yhteistyötÀ myös ensi vuonna. Q3 2023 (varmasti)-Q2 2024 (arvio) vÀlillÀ hankitut asiakkaat on jo maksettu takaisin eikÀ nÀiden vakuutusmaksuista siirry enÀÀ latiakaan GC:lle.
Totaalisesti kusessa? Hmm, ehkÀ, kuka tietÀÀ. Pohdi kuitenkin vielÀ sitÀ, ettÀ onko olemassa se mahdollisuus, ettÀ sun on vaikea kelata tÀtÀ tarinaa eteenpÀin ja miettiÀ, ettÀ mitÀ tapahtuu kun yhtiö skaalautuu ja kasvaa kokoonsa. Reppu on vielÀ liian raskas, koska skaalautuminen ei ole valmis.
NÀytÀ mulle se yhtiö, joka palvelee nÀin rajusti kasvanutta ylÀriviÀ ja asiakasmÀÀrÀÀ vain hieman kasvaneella opexilla.
Q3 2025 oli jÀlleen FCF+.
Kaikki viittaa siihen, ettÀ shorttipositioita ei ole vielÀ edes suljettu tÀssÀ nousussa.
Yhtiön kiinteÀt Sales & Marketing -kulut ovat n. $10M kuukaudessa. GC:n kautta investoivat nyt n. 45-50M/kvartaali asiakashankintaan. Niin kauan kuin LTV/CAC on niinkin hyvÀ kuin 3:1, toivon, ettÀ heittÀvÀt niin paljon dollareita siihen kuin mahdollista.
Valitettavasti nuo investoinnit tulevaisuuden kassavirtoihin valuvat suoraan ja heti tulosliuskaan, kun taas ansainta saadaan vasta vuosien mittaan / pÀÀstÀ. On tavallaan hassua, ettÀ kun autovalmistaja rakentaan tehtaan, jolla investoi tulevaisuuden kassavirtoihin, niin se luetaan capexiksi, eikÀ se rasita tulosta kuten CAC.
S&M kulujen prosentuaalinen osuus vakuutusmaksutuloista tulee laskemaan selvÀsti. On hyvÀ muistaa, ettÀ ensi vuonna liikevaihto tulee nousemaan esim. Q1 2026 >65% y/y.
Q3 2025 yli 50% uusista autoasiakkaista olivat $0 CAC:lla, eli myyty olemassaoleville asiakkaille. Auton GLR 76% = -16%-yksikköÀ y/y.
Eli strategian mukaan mennÀÀn: Markkinointipanostuksilla nuoret asiakkaat sisÀÀn pÀÀosin renters / pet -tuotteilla, ja sitten kasvetaan heidÀn mukana ja myydÀÀn autovakuutus. Ei tÀmÀ sen vaikeampaa ole. 
Premium per customer on $404, eli on helppo ymmÀrtÀÀ, ettÀ yhtiö on renters/pet -painotteinen, koska autovakuutuksen keskihinta Q3:lla oli $1964. Auton GLR alkaa olemaan skaalauskunnossa, joten myös PPC tulee nousemaan isosti tulevina vuosina.
Lemonade on melko harvinainen yhtiö siinÀ, ettÀ heitÀ ei kiinnosta lyhyen tÀhtÀimen tulos. He rakentavat painavaa ja merkittÀvÀÀ yhtiötÀ. Jos he ruuvaisivat kasvupanostukset minimiin tai nolliin, he olisivat jo ebitda+.
Kun homma aikanaan kÀÀntyy skaalautumisen myötÀ, uskon sen liikkeen olevan merkittÀvÀ ja nopea.
On tietysti tÀysin mahdollista, ettÀ olen vÀÀrÀssÀ. On totta, ettÀ Lemonade ei ole koskaan tehnyt tulosta ja jos yhtiöön sijoittaa, tÀytyy siihen tutustua huolella ja ymmÀrtÀÀ, mihin sijoittaa.
TÀmÀ ketju saattoi kuitenkin osoittaa sen, miten yhden lyhyen vuoden aikana todella moni karhujen FUD:sta on todettu vÀÀrÀksi. Eilen CFO sanoi, ettÀ he pystyvÀt kasvamaan nykyistÀ tahtia ilman lisÀpÀÀomaa. Valehtelevatko he?
KetkÀ olivat ÀÀnekkÀimmÀt Tesla-epÀilijÀt? Auto-alan asiantuntijat.
KetkÀ olivat ÀÀnekkÀimmÀt Netflix-epÀilijÀt? Elokuva- ja TV-alan asiantuntijat.
KetkÀ olivat ÀÀnekkÀimmÀt Amazon-epÀilijÀt? Tavaratalotoiminnan asiantuntijat.
KetkÀ olivat ÀÀnekkÀimmÀt iPhone-epÀilijÀt? Kaiken maailman IT-asiantuntijat.
KetkÀ ovat ÀÀnekkÀimmÀt Lemonade-epÀilijÀt? Jep.
KyllÀ, vakuutusalaan liittyy paljon regulaatiota ja pÀÀomavaatimuksia, joita teknologiayhtiöillÀ ei kasvun jarruna ole. Mutta kaikki tuntuu mahdottomalta silloin kun olet syvÀllÀ jonkun alan syövereissÀ. Jos joku tekee asioita eri tavalla kuin mihin sinÀ olet omassa laatikossasi tottunut, se tuntuu vÀÀrÀltÀ, huijaukselta, hÀmÀrÀltÀ tai mahdottomalta. Se on inhimillistÀ.
Ja muistakaa vielÀ:
Q4 2026 adj ebitda +
âHeh, mikĂ€ vakuutusyhtiö raportoi adj ebitdaa, tĂ€ysin turha mittariâ
Q4 2027 +nettotulos
âHeh, niillĂ€ kesti 12 vuotta pÀÀstĂ€ tĂ€hĂ€n pisteeseenâ
2028 â> 
âEn olisi kyl koskaan voinut uskoa, ettĂ€ Gen Z:n ostokĂ€yttĂ€ytyminen ja AI:n kehitystahti disruptoi vakuutusalaa nĂ€in nopeastiâ
(fiktiivinen tarina, ei sijoitussuositus)
Chaa. Writing on forums can sometimes be quite difficult, however, itâs not a competition about who is right and wrong â just arguing for and against things.
As for Lemonade: the companyâs underwriting result is still approximately 200 MUSD in the red â there has been very little improvement over the year. At some point, this will also materialize as cash. Insurance companiesâ cash flow statements usually have no analytical significance: premiums are paid first, losses later, and if there is growth, the cash flow should indeed be positive and the companyâs assets should grow significantly.
It is true that working in the industry makes one skeptical about certain things. Itâs worth remembering, however, that many have tried a model similar to Lemonade before, without success. The company does not have any special technology. The biggest problem with the business model itself is usually the costs of customer acquisition for (direct) sales, which are very challenging to push down. Lemonadeâs trend is still weak; sales and marketing expenses are still over 40% of earned premiums â for the model to truly start working, this must decrease to a level of ~20%.
As for the companyâs reported figures, these are notoriously bogus. If a company is unprofitable, its cash will dry upâŠsooner or later.
It is difficult for an outsider to get a proper picture of the company itself, because the company only reports the (self-developed) key figures of the parent holding company. The figures for insurance companies (i.e., risk carriers) are available, but, for example, the figures for various service companies (where the losses are) are not. (Note: at least in the last review, the results of the insurance companies are profitable due to transfer pricing. This is quite clever in itself, though not very transparentâŠso somewhere, cash is burning significantly).
As for the JV change, this should also be reflected somewhereâŠthe companyâs net insurance liabilities should increase considerably (as well as the capital requirement)âŠthe companyâs figures, however, do not mention solvency at all.
But whatâs the point of talking about it any further. Q4 figures will be reviewed again. It will be interesting to see what the future brings â good luck to the investors, of course (going forward). On a general level, I recommend digging deeper into Lemonadeâs figures and normalizing them to typical insurance industry standards.
Insurance companiesâ cash flow statements usually have no analytical significance: insurance premiums are paid first, losses later; if there is growth, the cash flow should indeed be positive and the companyâs assets should grow significantly.
Precisely. The cash remains stagnant as growing premium revenues, in a constant chase, offset subsequent compensations. Growth and the change in the JV agreement have increased and will continue to increase capital requirements, and that capital must be found somewhere if the results side is making a loss. Nothing is whispered about this in the interim reports, as you noted.
Thanks for the answer! Iâll try to explain it in simple terms, as itâs evident that you havenât yet delved deep into the companyâs figures. I hope these answers clarify the overall picture.
Thatâs true, but you surely noticed from my previous long message that the âtruthsâ thrown around on this forum have proven to be false.
Now youâre mixing up the terminology, arenât you? The underwriting result practically means the result of the insurance business before investment income and before other operating expenses: net earned premium â claims (incl. LAE) â insurance acquisition and maintenance costs. For Lemonade, this is best represented by the companyâs reported gross profit â and it has been positive and growing, not â200 M$. For example:
Q4/2024 gross profit $64M, margin 43%; full FY2024 gross profit $167M (+98% y/y).
Q2/2025 gross profit $64M, margin 39%.
Q3/2025 gross profit $80M, margin 41%.
The companyâs net result will be approximately -170M this year.
Are you aware of what largely causes this and how General Catalystâs funded Synthetic Agents, i.e., growth financing, works?
In 2024, Lemonade spent $121.5M on growth. General Catalyst funded 80% of this, Lemonade itself 20%.
In 2025, Lemonade will spend $170M on growth. GC funds 80%, the company itself 20%. Almost $50M more than in 2024!
Also, the portion funded by GC, which does not come from Lemonadeâs cash, goes directly and immediately to the income statement as a negative. The company is thus buying future cash flows and revenues, i.e., customers (in Q3, LTV / CAC was already almost 5), for which the expense arises immediately, but the revenues only later. This is the most important thing to understand.
Thatâs why the net result lags, even though cash flow is positive.
The company repays the financing for the monthly customer cohorts funded by GC from the premiums of those customer cohorts + 16% annual interest. These are included in the G&A on the income statement. Once the financing is fully repaid, 100% of the premiums come to Lemonade. This wheel is thus spinning with increasing power every month.
Customers acquired between July 2023 and spring 2024 have already been fully repaid.
Who has tried this model in the AI era? I argue that no company has built its entire business around and on artificial intelligence, every function of the organization.
âThe company has no special technology.â OK. You should watch Investor Day 2024, itâs on YouTube, if you really want to see whatâs being done under the hood.
As proof of special technology, I believe itâs enough that OpenAI has repeatedly highlighted Lemonadeâs technological work in its posts and that Metaâs Prashant Ratanchandani (âVice President of Engineering for AI Products, leads the engineering teams responsible for building Metaâs Generative AI productsâ) just joined the companyâs board. I donât believe there can be any other reason for this than special technology. Lemonade is, after all, only a small company worth 5 billion.
Special technology is also indicated by LAE, which is only 7%, and when the business doubles, the company said this would decrease to 3.5%. For traditional US companies, this is >9%. Special technology is indicated by the fact that the companyâs headcount decreased Y/Y in Q3, even though 500,000 new customers were added.
I think this looks quite good, considering that PPC has also been growing all the time.
You have overlooked here that the figure is hugely affected by the fact that the share transferred to reinsurers has been 55%, and from July 1, 2025, it will be 20% for new/renewed contracts.
From July 1, 2026, it will be only 20% for all contracts.
If NEP was about $140M in your Excel for Q3 2025, it will be about $280M in Q3 2026. This growth is so sharp precisely due to the quota share change. And when about $50M is spent on marketing in Q3 2026, you will notice that the share of marketing in earned premium is only 18%.
Provide examples of bogus figures here, and we can discuss them? I understand that the lack of CR and the companyâs use of adjusted figures cause irritation among traditional insurance companies.
âIf a company is unprofitable⊠its cash will dry up.â In this case, it seems you might not have understood how General Catalystâs financing model works.
Lemonadeâs RBC ratios clearly exceeded the minimum requirements at the end of 2024 (5-6 times):
Lemonade Insurance Company: 577%
Metromile Insurance Company: 525%
The JV change was possible because loss ratios have continued their sharp decline.
Yes, it is often mentioned.
"..And finally, all else equal, less quota share increases regulatory capital needs. However, with an improved loss ratio and the expanded use of our wholly owned captive, we are able to offset these pressures such that there is no material change in our capital planning.â (CFO Tim Bixby, Q2 2025 earnings call)
Hopefully, these comments helped you understand beyond a superficial glance?
Itâs worth getting to know Ping An company better, for example. See
Driving Innovation with Generative AI in Healthcare & Finance
However, itâs on a completely different level compared to LemonadeâŠlarge insurance companies have their own AI projects - Lemonade has nothing special. Everyone should do their own analyses and draw their own conclusions - a significant short position speaks for itself.
You are right that China has been ahead of Western countries in many technological fields for a very long time. Would there be any examples from Lemonadeâs competitive landscape, i.e., America or Europe?
I think that the large short position is largely due to superficial analyses being made of the company, such as on this forum, and because Lemonadeâs net result is heavily loss-making.
In investing, itâs quite dangerous to think that a large short volume would indicate something about whether a company is a good investment. You may recall examples from history. You might miss out on big returns.
I would gladly continue the discussion with you about the points I made in my previous posts!
Even before Lemonade was founded, large insurance companies started establishing their own fully digital companies. Baloise, for example, founded a company called âFridayâ, and Munich RE founded âNexibleâ. Both companies were shut down in 2025 due to heavy losses (you can ask ChatGPT about the history of these companies, for example).
One of the biggest problems has been customer acquisition costs, which cannot be driven down enough to make the business profitable â the same problem that still plagues Lemonade. Even the âold onesâ have tried these new models â without success. How much AI is ahead of the 2020 IT stack and process automation, itâs hard to say. AI itself is being tried to be integrated into processes instead of a complete re-establishment, where the old players have a big data advantage. Ping An is, at least for me, the most advanced in the market in many areas â see, for example, Ping An - AI Smart Instant Claims - One Connect. So yes, these AI applications are being developed everywhere with big money.
Short positions usually come from the institutional side, where there is a lot of analytical capacity. Itâs true that it doesnât prove anything in itself, but a large amount indicates that there is enough skepticism â towards valuations or the business model.
Indeed. If you invested in Lemonade at the beginning of the year, youâve made great returns. Will it work out in the end â time will tell. For me, the biggest red flags are customer acquisition costs (unprofitability of insurance operations) and the lack of transparency in reporting (see, for example, Ping An, the company is perhaps the toughest AI insurer in the field, but investment presentations are very insurance-like). I tried at some point to model Lemonade on a bottom-up basis (i.e., through the holding companyâs subsidiaries â practically without looking at the companyâs own reported figures)⊠this is hopeless because itâs not possible to get income statements and balance sheets for service companies.
But, but. Letâs see how it progresses.
Traditional insurance companies have dozens of legacy systems that donât communicate with each other. Building AI models on top of these is expensive and slow, and often not economically viable. A much more realistic option would be to move data to a platform like Palantir and streamline operations through it. However, the insurance industry is old and conservative, so I am skeptical about how quickly it can adapt to the change brought by AI.
Lemonade was founded in 2015, while the companies you mentioned were founded only after that. If you look at it through ChatGPT, it notes when Lemonade entered the European market when comparing it to European counterparts. These are also not very good benchmarks for Lemonade, as they did not use LLM models. They were more traditional digital insurance brands, and their unprofitability was more due to their failure to scale down costs through mobile, web, and automation solutions.
Traditional insurance companies have a lot of data, but data itself is not valuable unless decisions can be made based on it. This is where LLM models are particularly useful, as they can quickly sift through vast amounts of data and highlight essential information. Your claim that Lemonade has no technological advantage is incorrect, as legacy companies have not yet started using LLM models in decision-making.
Lemonadeâs business is difficult to grasp, and I would argue that many short sellers do not look at the matter beyond Excel numbers and ChatGPT answers. It is good to remember that the insurtech model has a strong technical dimension, meaning we are at the forefront of something new, where AI-based claims automation and customer acquisition logic are changing the entire operation. This overall picture is easily missed when Lemonade is compared to counterparts that do not yet have similar technology in use.
Now I have to chime in, as for once we are discussing something I can claim to know something about:
Traditional insurance companies have dozens of legacy systems that donât communicate with each other. Building AI models on top of these is expensive and slow, and often not economically viable. A much more realistic option would be to move data to a platform like Palantir and streamline operations through it.
That legacy system is not problematic from a data perspective, but rather from the perspective of operational capabilities. That was already tackled over a decade ago when Data Warehousing was developed, and now Snowflake, Databricks, and ClickHouse are already leading the third generation of Data Warehousing (The second was Lakehouse, and Databricks at least calls the third Data Intelligence Platform).
However, the insurance industry is old and conservative, so I am skeptical about how quickly it can adapt to the change brought by AI.
This doesnât really relate to the insurance industry itself, but traditional machine learning is becoming a de facto part of business in almost every sector. For example, IF runs fraud detection, speech recognition, and part of claims processing with traditional machine learning models, even though they also have honest-to-goodness mainframes.
Here, LLM models are particularly useful, as they can quickly go through large masses of data and highlight relevant information.
This is true, but a traditional data engineer / data scientist can dig up that information in the same way. Usually, the information brought up by an LLM still needs to be validated using traditional methods. It doesnât take much longer for an expert to write a query vs. a prompt, and the data still had to be processed and cleaned. LLMs primarily facilitate experiments for business-side people.
Your claim that Lemonade has no technological advantage is incorrect, as legacy companies have not yet moved to using LLM models in decision-making.
An LLM isnât very good at making decisions; it merely tries to predict the next words. Chatbots like Gemini etc. can, of course, also handle traditional data processing, but they are implemented with an agent framework like LangChain, which allows the main agent to be given sub-agents that can be, for example, traditional Linear Regression models or even just rule-based SQL commands. The LLM then merely binds the results into a verbal form, but it doesnât make actual decisions.
Finally, a couple of general observations.
Does a high degree of automation attract high-CLV customers, or does it only gather low-margin customers? What if high-CLV customers prefer real customer service representatives? Examples of this from the Finnish banking scene include Handelsbanken and Ă landsbanken.
Hereâs a Forbes article, the title of which contains everything essential: MIT Says 95% Of Enterprise AI Fail- Hereâs What The 5% Are Doing Right
For example, my own employer wants to invest heavily in AI, but most of the âAIâ runs more cost-effectively on traditional machine learning algorithms or precisely rule-based data processing. An LLM can be used, for example, to generate 100 different greetings for a customer app.
In the big picture, business focused on LLMs is not very scalable, as the number of calls sent to the language model increases with the growing customer base, whereas traditional machine learning inherently produces better results as the amount of data (e.g., number of claims) increases.
The fact that traditional âlegacy companiesâ do not use LLMs in their customer interface does not mean that the company does not otherwise utilize LLMs as part of employeesâ daily work. The fact that âlegacy companiesâ do not use LLMs does not mean that the company has not tried them. Perhaps it concluded that sufficient cost benefits were not available in those cases, but it does not need to cling to the AI hype to secure its stock price and future.
Traditional insurance companies have a lot of data, but data itself is not valuable unless decisions can be made based on it.
This is more than true, but letâs emphasize that AI built on non-existent data is also worthless. Generally, data-driven management still surprisingly often means something like economistsâ Excel sheets. An example would be that green car dealer.
Itâs good to remember that the insurtech model has a strong technical dimension, meaning we are at the forefront of something new, where AI-based claims automation and customer acquisition logic will change the entire operation.
Isnât customer acquisition quite generally âAIâ-based already, if we look at online advertising, for example? On the other hand, at least in Europe, legislation, if I remember correctly, restricts robot calls, so Lenni Lemonade cannot start calling in fluent Savo dialect; that is still human territory.
Despite all the hype, AI is no silver bullet, but rather an economistâs wet dream of one. Building a reliable and high-quality AI solution requires just as much work as building a high-quality reporting solution, and building a data pipeline is certainly far from sexy.
Currently, my opinion is that language models are good precisely for supporting the work of experts. They can help a skilled coder write functions or correct errors. They can help a marketer write lively text and correct typos. The internet and automatic data processing were an industrial revolution, but language models are hardly that, even though they have undeniable advantages. Still, they are just one tool in an expertâs toolbox.
I emphasize again that I have not familiarized myself with the company but am commenting generally on the arguments presented in the thread about the omnipotence of AI. I also emphasize that the development of language models is currently supported by enormous resources, so the situation may change very quickly.
However, be careful. The 2020s have seen a huge number of âdisruptors,â but how many of them have actually been a commercial success?
Itâs not that the problem hasnât already been partially tackled, but rather that the solutions you mentioned only reduce legacy challenges â they donât eliminate them entirely. The real problem still relates to data fragmentation, quality, and integration costs. It is precisely to these that platforms like Palantir bring added value, as they focus on supporting operational decision-making and managing workflow integrations. Lemonade stands out in this competition with its structural AI-first model, which makes it agile and cost-effective. Legacy companies can indeed âmitigate problemsâ with data warehouse solutions, but they wonât achieve the same speed and customer experience without massive modernization â or by becoming a customer of a platform like Palantir.
I donât disagree with this, but the example might be a bit simplistic. Modern LLM agent developments (e.g., UiPathâs agentic automation solutions) enable the model to guide the decision-making process, choose different paths, and orchestrate traditional ML models.
Regarding scalability, I agree: traditional ML is still cost-effective in many confined tasks, as you mentioned. However, LLM brings added value precisely in complex and customer-facing tasks that rule-based logic does not cover.
I understand the skepticism towards AI, but it is equally a mistake to underestimate the evolving role of LLM technology in supporting decision-making and improving customer experience â precisely what Lemonade is investing in.
AI itself is no silver bullet, and many new AI players in various fields tend to build all sorts of world-improving contraptions based on it.
In the insurance industry, however, Lemonade is a good example of how technology can transform the entire business model when applied correctly and leveraging its strengths.
Time will tell how this model works without legacy systems, with an AI-first architecture: a chatbot handles customer acquisition, underwriting is based on real-time data, and claims processing is largely automated. This makes the company agile and cost-effective compared to traditional legacy companies, for whom AI is often just an additional layer on top of old processes.
ValueByMarkus has written about Lemonadeâs Q3. 
Lemonadeâs Q3 was a strong indication that the company is delivering exactly what its investment story promises: the ability to grow, improve profitability, and leverage automation in a way that is not commonplace in the traditional insurance industry. In Force Premium (IFP) growth accelerated for the eighth consecutive quarter, and at the same time, the Gross Loss Ratio (GLR) continued its decline. For an insurance company, this combination is a rare treat â usually, accelerating growth leads to weaker loss ratios,
Could someone explain to a dummy, how does Lemonade innovate the insurance market? Despite looking into it, I still havenât understood what Lemonade does that, for example, IF or some other traditional insurance company couldnât do? Fundamentally, the argument âwe do the same as others, but we use AI and automationâ doesnât seem very convincing in the long run. Digitization and automation of services and processes are probably not the hardest thing in the insurance business?
Yes, it seems surprisingly difficult. Geico tried for over 10 years to move its operations to the cloud, but eventually gave up because it was just too difficult, and returned to physical servers.
At Berkshireâs annual meeting two years ago, Geicoâs CEO stated that they have 600 systems that donât communicate with each other.
Large companies in the US are steadily recruiting COBOL experts to keep their systems running.
Modernizing old systems is not easy or necessarily even possible.
However, an even greater challenge, in my opinion, is culture. If the insurance business relies on the work done by insurance salespeople, one can be quite sure that there is no enthusiasm there to make their own work redundant.
A lot has been covered in this thread. I also recommend, for example, the X LMND community and paperbagâs YouTube videos.
One example https://x.com/shai_wininger/status/1988226889832689718
The last time I made a claim to a domestic insurance company, it took about two weeks to get a decision. In this example, processing took about 3 seconds.
Lemonade now available integrated with Tesla.
The price is dynamically determined by mileage (pay per mile) and driving style (car data).
https://x.com/shai_wininger/status/1996272516793241993?s=46&t=qDMR_G-hy9voJ-cjq9_ZWA
Itâs also worth reading the bear cases on Lemonade. For example, @AP_1981 raised good warning flags about a month ago.
The reported figures are confusing, and itâs difficult or impossible to extract meaningful financial figures from them. The company will continue to generate losses for some time, the multiples are high, and the number of shares is constantly increasing, in my opinion, through a rather large stock compensation program.
On the other hand, financial figures are constantly improving, and the company is growing rapidly, so it certainly has the potential to capture a significant share from traditional companies by being clearly more dynamic. The market capitalization is only about 5B USD, compared to, for example, Progressiveâs 134B USD.
I myself am bullish and involved, wondering how the story will unfold. I am neither selling nor buying at the current price.
OP has recently had an interesting recruitment open in the insurance field.
In my opinion, the description clearly shows how out of touch legacy insurance companies are with the future of the industry, especially compared to, for example, Lemonade. The announcement consists mostly of platitudes, and the required skill level is straight out of the 90s. The position is not looking for a strong technical visionary for the future, which OP Pohjola undoubtedly needs to build the whole thing from scratch towards autonomy.
IF, OP, Fennia, Geico, etc., will still be caught with their pants down. The silver lining for domestic companies might be that we are in a side market here, so business might stay alive a little longer. But eventually, Lemonade or another similar progressive company will also take over the Nordic countries.
Here is the announcement in case itâs not visible soon.
[spoiler]Are you a deep expert in the non-life insurance sector, inspired by the possibilities of technology as part of customer-centric business? Can you successfully and collaboratively drive change?
We are looking for a Head of Pricing and Underwriting Analytics for Pohjola Insurance to be responsible for leading and developing pricing and underwriting analytics as part of achieving Pohjola Insuranceâs strategic goals.
About the Role
In this role, you will get to build a superior customer experience, implement Pohjola Insuranceâs strategy, and lead your area of responsibility in accordance with business objectives.
In this role, you will lead Pohjola Insuranceâs Pricing and Underwriting analytics entity, which comprises two teams and approximately 30 people. You will have direct supervisory responsibility for the operational supervisors of these teams, and you will ensure that the teams have the necessary expertise, resources, and prerequisites to succeed in their set goals.
You will create the target state and roadmap for Pricing and Underwriting analytics together with the Products and Services and Business Analytics areas of responsibility, and you will be responsible for its progress in accordance with agreed operating models, methods, and technological solutions. In your role, you will also be responsible for cross-business collaboration regarding analytics, and you will consider various stakeholders at the OP Pohjola level.
What kind of expertise are we looking for?
We require a deep understanding of non-life insurance business and experience working in a customer-centric business model. In addition, we expect experience and proven track record in analytics and database expertise as part of data-driven business understanding.
You are capable not only of developing but also of renewing operations, expertise, and capabilities and implementing them with a bold and inspiring approach. You have the skill to lead, plan, and organize entities and collaboration. You possess good communication, collaboration, and networking skills and actively participate in discussions around your area of responsibility. You continuously develop your own expertise and common best practices and are energized by teamwork and achieving results.
We require a suitable higher education degree and fluent English and Finnish language skills for this position.
What we offer
In this role, you will get to challenge yourself with diverse tasks at Pohjola Insurance and will have a vantage point into insurance services for private and corporate customers. Come create the best experience for our customers with us!
Our operations emphasize a strong goal-setting culture, an approach that strongly empowers our personnel and emphasizes self- and team-direction, and a coaching leadership style. In your role, you are key to setting the direction and ensuring that the Pricing and Underwriting analytics teams have every opportunity to succeed.
In counterbalance to responsible work, we take care of your well-being at work and future-oriented work arrangements. We offer you a stable environment where responsibility, human-centricity, and succeeding together are key values.
You will have access to our diverse employee benefits, which include an employee fund as well as staff-rate banking, insurance, and loan services. We also offer comprehensive occupational health care, extensive medical expense insurance, and full-time accident insurance. We invest in work-life balance and flexibility for well-being at work, and by offering you lunch, sports, and cultural benefits, as well as the opportunity for affordable holidays in OP Pohjolaâs holiday properties.[/spoiler]
It sounds like Risto is talking about Lemonade in this few-minute clip?
AI-native vs Legacy company. In my opinion, that perfectly describes Lemonadeâs head start and hopefully opens up a perspective that LMND bulls have @Eicca.
