Carpe Data gathers and aggregates data from from Yelp, Osha, Glassdoor, crime-databases, Twitter, and other data sources. Carpe cleans the aggregated data, builds models on it and sells it to the insurance companies. A good example is how Carpe Data serves commercial insurance. Let's take a restaurant as an example. Carpe analyses the business’s website, checks if it has a happy hour, a deep fryer, hours of operation and crunches more data points, and delivers a clean data set for the insurers to consume.
Max recognizes several value offers:
- Pre-filling applications that helps pre-qualifications
- Qualification validation based on multi sourced data
- Insurance/risk score for small business that allows a better decision making when adding a business to the book
- Something in the data points
- Provide data to underwriters
Q: How much of this is, or will, take the commercial underwriter out of the picture? How much of it is automated?
A: If you ask any carrier about their commercial, or small commercial, they will say that it is automated. None of them automate, except for the really really small stuff and then they don’t validate the data. They are passing data through rules and hope for the best. I think that they look at the disruption in the auto insurance and think how they can take it to commercial.
Carriers should take a look at the operation of commercial underwriting and its input. I ask “how many underwriters use Google street view and how many visit the property?”.
Q: How is the investor echo system reacting to this new trend in InsurTech?
A: It is insane. In the past year there were legit 5-6 conferences. They try to apply what they know from SaaS into insurance. The prize is big, but the sales cycle is very long and there is no predicting it. It is hard, unpredictable and a turn-off. I told investors that “if you have this entrepreneur that had a success and now he wants to import it to insurance... He is wrong, he will fail”. Many investors see it as the last “white space”. Others focus on the distribution like Lemonade, Trov, Quilt, Goji, Ladder Life and more. Most of them are basically agents that re-sell insurance and try to put on it a better UI.
I don’t think that insurance has a branding problem, and I don’t think that people hate their insurance. I think that people like Farmers, All-State and Progressive. I don’t think that the distribution will be the great disruptor of the insurance industry. But, most of the money is there.
Q: What about global insurance for professionals who travel and work from various countries?
A: I don’t know. There will be more and more new insurance products that will need to compensate for the shrinking of the auto insurance in the near future. Pet insurance is a great example for a booming product that surfaced in the recent years.
Q: How your customers in the IT departments respond to this change? (data services)
A: Some actuary groups try to do a little bit from this and a little bit of that. They don’t want to build their own API connector. They don’t want to tap and pull data from 50 new APIs. They will have couple of data scientists that will do it as an experiment, but that is where they’ll stop. They want, and like, to buy data. For example, Lexis sells data to insurance companies in billions of dollars. The IT departments care about data, and getting the data.
Q: Do you collect data regarding cyber security?
A: That’s a great question. One of the product ideas that I thought about, which we don’t do, is specifically for cyber security and cyber insurance. There are companies that sell data about IT stacks and they understand what the company is using internally and externally. Are they using Salesforce? What kind of firewall? Or are they using AWS? And as far as I can tell, cyber insurance, base their price on non-sense. There is no bases or anything. They don’t check if you have the latest patch, or if you are running Ubuntu, or what ever. They don’t ask these questions. They ask “what is your revenue -- ahh the price is $2,000 a year”. I think it is a good opportunity.
Q: What is your recommendation to the entrepreneurs in the audience?
A: My recommendation to entrepreneurs who want to go to insurTech is go and work for an insurance company. In one minute you will recognize a thousand opportunities to disrupt. You need to figure out what are they doing wrong. Then, you want to learn as much as you can how it works and then you’ll go and build the right thing and sell it back. Insurance carriers don’t abuse vendors, they have a slow procurement cycle that you need to know and accept. They are not horrible to work with, they are just rational and risk averse. I will suggest to get experience with an insurance company. I think that the best thing is to go work for one. It will be a huge differentiator to start a company that tries to serve insurance carriers after actually working for an insurance company. There are so many startups out there that don’t know anything about insurance.
Q: Why is the sales cycle so long?
A: People will often say that the insurance companies are risk averse. But, also, the deals are very big. The size of the business rather than the aspect of it is the factor that adds to the time length. You need to learn and understand the process. It takes 3 months to get the business unit engaged, once they are engaged, they will send you over to IT. IT will look at it for another 3-4 months for validation. Then procurement will process it for 3 months. Once you get the work order, the company will need to train everyone and that will take another 3 months. And that doesn’t include an internal campaign that takes 3 months. Going back to the opportunity, if you have the patience and the understanding of the process you have an advantage.
Q: How the investors react to this reality?
A: Some understand, others don’t. They will ask “I’ll bring these superstar salespersons, do you think they can accelerate the sales process?”. The answer is “no, it is what it is”.
Q: How the millennial and the gen-Y consume insurance?
A: They don’t buy life insurance! This is a big thing. They were not taught about it at school, they don’t know about it. Employers used to provide it - not anymore. Government jobs used to provide it - not anymore. I do think that young people will not buy car insurance. That is a tipping point. The first part of this tipping point will be a dramatic differences between premiums for cars, which are safe thanks to auto pilot technology and cars without the technology. The second part is that Tesla, or Mercedes, or another self driving car will kill somebody and his family will sue the insurance company. The insurance company, then, will sue Tesla. Then everything will change. Tesla will say - you buy a car and the insurance from us, you will get everything from Tesla. They should do that because they are already on the hook for liability.
BTW - what is safer? A 16 years old driving a car or an Uber driver?
Q: We are looking at all these data points for underwriting 1 to 1 scale. Do you think it will apply for a risk pool? Do you think these techniques will scale?
A: Probably. Look at group health application and compensation in general. I am better at 1:1 ratio. Carriers use us to buy a book of business, because we can look at the risk differently. So a book will be a bundle of risks and we can look at how much a package of risk is worth. And we can point to several policies that they should not buy.
Q: Are you looking into reinsurance?
A: The re-insurance are lighting up the investment world. They have invested hundreds of millions of dollars over the past couple of years. They love innovation and they wish to run pilots. If that pilot is successful they will incentivize their carrier partners to use your technology. It is too soon to tell. They are trying.
Q: Are they fishing?
A: Maybe. There is a rumor that they see all the new digital insurers and want to go in for themselves.