We met up with Robert Malooly, CEO of Claim Maps. Claim Maps is based in Olympia, WA and provides software in the workers compensation space.
Krish: Hi Bob, pleasure to meet you. Thanks for taking the time to talk to us and taking the time to drive from Olympia.
Bob: Oh Sure. Pleasure to meet you as well.
Krish: Bob, Can you tell us about some highlights of your career before Claim Maps?
Bob: The reason I got into workers compensation business, which directly led to me starting Claim Maps was that I had worked in the Illinois unemployment insurance program. The program at that time went bankrupt, the state owed the Federal govt $2.5 billion dollars. By default, I wound up handling all the loans and suggested building an actuarial unit to guide the restructuring of the program and put the plan on a sound financial footing. I offered to recruit a fellow from the Society of Actuaries to do the work. When the plan was presented to the Governor, he said it is great, but there is no time to establish actuarial positions, recruit and hire, he asked can you do it yourself? That led me to the actuarial business. In the process we built very very large models of the system and worked with business and labor to get agreement on reforms, rewrote the tax code and made a lot of other changes. It worked.
Sometime after that the Governor asked me if I wanted to be Chairman of the Industrial Commission. They needed someone who was politically neutral to implement a package of legislative reforms. The Commission is the administrative court and regulatory body that handles workers compensation. What was supposed to be a 9 month stint turned out to be 6 years. That was my introduction to the workers comp
After that I was recruited to a startup that was in the insurance analytics business by Aon, a very well-funded operation that came to a rather tragic close due to the death of a key backer in a plane crash. I was then recruited to a second startup in the same space, which then sold to ISO. I then worked for ISO as Director of New Products for workers compensation. They are a huge property and casualty insurance data collection and rating organization. From there, I got recruited into the State of Washington overseeing the Worker’s compensation fund at Department of Labor and Industries, the 8th largest workers’ compensation insurer in the country.
All through my experience above, I used visual explanations to help people understand very complex and difficult problems. The visual approach led to much greater understanding.
Krish: Let me back up there a little bit and ask you to explain Worker’s compensation in simple terms for our readers. Not many people in the tech industry are aware of it.
Bob: Lot of tech people have no idea of it. It is an obligation for employers to carry insurance should any worker get injured on the job. For construction, logging, high hazard jobs, it is an extraordinarily expensive thing to carry and could be equivalent to 50% of an hourly wage in some extraordinarily risky areas. For tech people, carpal tunnel is probably the biggest exposure, so it is not a big deal. It is about $70b in annual costs to American businesses, $40b of that in direct insurance premiums, the rest in high deductible expenses and self-insurance costs.
Krish: Do all companies need to carry workers compensation insurance?
Bob: It depends on the state. Yes, anything bigger than a sole proprietorship has to carry comp insurance. In some countries, the insurance is not offered by private companies, but offered as state social programs, in Germany for example. In United States, it is mostly private.
Krish: These days, the buzz word in tech is “disruption”. Everyone wants to disrupt traditional methods of doing business that have been in place for last 50 or so years. Given the legacy in worker’s comp, why choose this as an area for a startup?
Bob: The primary reason is that I understand it very well. I am also well known in this area. I participate in a lot of national organizations and have regulated self-insurance in both Illinois and Washington. I am also familiar with the problems that companies have in workers compensation. It is a problem that really needs a solution. What I have seen in my experience is that a lot of people who are maximizing components of the system do so to the detriment of overall costs for employers and outcomes for injured workers. A CEO of a large TPA (third party claims administrator) told me he did not care about saving costs for his customers. He only cared about money that would go to his bottom line. As a consequence of that, you wind up with people doing things in the context of their immediate concern, that makes a lot of sense for the individual vendor. It is Herbert Simon, the Nobel Prize winning economist, who called this bounded rationality. That is where people’s thinking is constrained by a narrow boundary. Within that boundary, everything makes sense. But as soon as you step outside of that boundary, you get a completely different perspective and can see waste and bad results everywhere. Claim-Maps has stepped outside the boundary to look at the system as a whole. We see wasted expense and bad results for injured workers. And, we know how to fix it. That is disruption.
Claim Maps is the only way that many employers get a chance to view both ultimate outcome and total costs and can see how to minimize costs and produce best outcome. The way it is now, the data is in small chunks, kind of spread out and each entity is maximizing their own bottom line at the expense of the employer and the injured worker.
Krish: so, this is an area ripe for disruption?
Bob: Yes. If you talk to employers about workers comp, they all hate it. A lot of bad things happen unnecessarily. It is a big big expense for them. When we show our product to claims people and employers, they literally jump out of their chair. They get so excited because it is the first time they have been able to see everything in one place.
Krish: You founded Claim Maps a few years ago. Given your long career with a sprinkling of startup experience in between, I wonder about your drive to create a new company after that long career. Talk to us about that.
Bob: I have turned around a couple of governmental organizations and startups are a lot like turnarounds. You are dealing with huge problems, with a lot of uncertainty that many people would not want to take on. I have successfully dealt with a lot of messy problems in my career, situations where they need someone to turn it around. Running a startup is not all that different from what I have done in my career. Though I was not an expert in actuarial services, I understood the problem well; I was able to recruit graduate students from UChicago and Northwestern to help build the models after I had done the first one. These kids were so smart that they were insulted if I bought them a simple problem. That’s what a startup is too – you have got a really challenging problem, you have a clear goal you want to get to, but the path is unclear. You have to figure it out, navigate the uncertainty, attract talent to help you and produce value at the end of the day.
Krish: The work you did in various organizations seemed like a trial run before Claim Maps…
Bob: I feel like everything I have done has led to this point. Yes it seemed like it was a natural next step. Some of my peers think I am crazy. But, I know the market for this product very well. We have no problem getting to see decision makers. We are dealing with enterprise sales, which is not fast. So, we want our investors to be patient. There are lots of people in large organizations that can kill a deal. People responsible for the area say they want it, but IT people say no or the budget people veto it.
Krish: Hold on to that. We will come to that in a minute. How did you adjust to the startup world? What was easy and what was challenging?
Bob: Well the ideas were really easy. I understood the problem very well. I am using techniques that have proved successful in the past to solve these kinds of problems. That aspect is really fun.
The biggest challenge is to get people (customers) to understand. Claim-Maps is the first of a new class of software; it is different from straight quantitative or predictive analysis. It is goal focused software. Getting the customers to understand that this is an opportunity to see both the big picture and the detail in a way that allows you come away with a very quick and sophisticated understanding of complex events. We haven’t built everything yet. We built the basic structure where we have taken the transaction level pieces of data. We literally have everything that happened for 281,000 claims. Taking that information and allowing one to see at high level, get some understanding of general metrics associated with that and then being able to dive down into the components that are driving those numbers. Usually it is 20% of events driving everything. In workers comp, it is usually 5-6%
Krish: Why is that difficult? Claims adjustors are data driven people. The ability to show historical events and then the ability to drill down, is that a new paradigm for them?
Bob: Yes it is. Absolutely. If you take the volume of work that a claims adjustor contends with, there are about typically 200-300 claims assigned to each adjuster with all kinds of activity with financial decisions, legal decisions and medical decisions all coming together at once. The Claims adjustor is expected to coordinate all of that and they are stuck with reading text and trying to interpret notes for 200-300 claims all at the same time.
Krish: Let me rephrase. If this work was started in a large organization, would they have been able to convince customers easily? Why is this unique to your startup?
Bob: The people who deal with the claims problem directly understand our product immediately. It is the challenge of convincing people in the financial and the IT organizations of our customers. Some of them are not empathetic to the problems of the claims adjustors, case managers and supervisors. They also don’t appreciate, in some cases about the hundreds of millions of dollars going out the door unnecessarily and producing bad results. In an enterprise sale like this, we have to get all the components together at once and make them go forward.
Krish: So, the enterprise sales aspect is challenging from startup perspective?
Bob: Sure. We are dealing with Fortune 500 companies. Our first company said they don’t want to work with a startup and even after telling us they liked our solution, they said they preferred to do business with an established company. However, after surveying the market, they could not find anything close and agreed to deal with us.
Krish: Anything else that stuck you as a challenge as a startup?
Bob: Nothing else really. I was used to hiring talented people, managing budgets and dealing with uncertainty. Biggest challenge is customer awareness and then finding the early adopters.
Krish: What would you recommend potential entrepreneurs? Get some domain expertise before starting a company or just go for it when you are inclined to?
Bob: Tough question. If you are going to an area where there is lot of tradition, like insurance and if you don’t have credibility with your customers by virtue of your previous experience, they won’t even talk to you. If you are going to something brand new, not bound by tradition and you have an idea, just go for it.
Krish: So, a partnership/advisory/mentorship between a new entrepreneur and an industry expert won’t work in such traditional areas?
Bob: If someone had come to me with a proposal for a business like Claim-Maps and asked if I would join in and make a successful enterprise, I would absolutely have done that. If someone had come forward with a potential solution with this problem I had been dealing with, I would have said – Let’s go!
Krish: What is the value proposition of Claim Maps? As I understand it, Claim Maps provides a visualization solution for workers compensation claims. Is it just a new way of looking at data?
Bob: Let me take an example. We have a Doc Compare tool that is in our roadmap. Let us say two doctors have identical low scores in a scorecard. The recommendation would be to throw both doctors out of your physician network. However, using Claim Maps tools, one would be able to drill down and see that Doctor 1 has been doing unnecessary surgeries, but Doctor 2 had been taking very difficult cases that other physicians had messed up and making progress and getting interesting positive results. Using Claim-Maps, you would elect to keep Doctor2 in the network. Limiting the analysis to just numbers is problematic. Humans are very good at pattern recognition. We don’t give them opportunities to see patterns in complex events in an easy fashion. Workers comp events are tremendously complicated involving lots of financial, medical and legal disputes. The average comp claim in California, for example costs $60k. We don’t give the folks who are responsible for controlling these expenses a decent set of tools to use. That’s what Claim Maps does. We give our users an entirely new set of tools so they don’t need to spend three days reading a claim file. You can see a visual representation and a pattern jumps out at you. You notice something unusual and you can drill down and not have to spend days looking at a claim file.
Krish: Isn’t this an area that lends itself well to machine learning? Let the machine identify patterns and either make decisions or send you alerts?
Bob: We already raise flags alerts based on some business rules and limited analytics. We are going to build a lot of capability in that area. May be long after driverless cars, we will have a system capable of making decisions. Until then, we are going rely on humans. Humans empowered by very effective technology like Claim-Maps to bring the best of machine and human technology to the problem.
Krish: Humans are going to make the decisions. Humans don’t have a lot of time to know what to look for. I see this as a signal to noise ratio problem. Does Claim Maps improve the signal amidst all the data noise in claim data?
Bob: Yes absolutely. As one example, we are going to use some sophisticated analytics to build our matching engine so we can have groups of similar claims compared to. How close is a claim compared to a typical back claim? Are there important differences or trivial differences? We have a cartoonist who designs our icons and I have asked him to design a train wreck icon. There is a pretty good wine called Train wreck (laughter). One of the problems with machine learning is competing algorithms running in the background to analyze data and let’s assume on algorithm says yes and the other says no. If one is 3 points better saying Yes than one saying No, which one is the machine going to go with? We want to have competing analytics running in the background so that when each tells a different story, a train wreck icon (jokingly) appears in the timeline and alerts an adjustor to look at what the problem really is. Machine learning is wonderful stuff that we are going to spend a lot of money building in, but it usually results in overconfidence on the part of the human doing the building. You are going to think you are better at machine learning than you really are, if you are the human designing learning systems. As an example I have seen this with surgeons, who have been told up until they become a surgeon that they are #1. They were #1 in kindergarten, and #1 all the way through medical school. Then when they do finally become a surgeon, they are shocked when they realize that among surgeons they are just an average surgeon. Claim-Maps is being designed to detect common human reasoning errors, overconfidence is just one of them.
Krish: Who is using Claim Maps today? What has their response been?
Bob: Right now, our first customer is Marriott. They saw a power point and they wanted it. After looking at everything else in the market, they asked for a contract and wrote us a cheque. That was 2.5 years ago. However, Marriott has been a blessing and a curse. Great name, great reputation in the workers comp space.
We are dealing with a Fortune 500 company and it has been challenging. Their IT staff is more concerned keeping their reservation system up than dealing with Claim Maps. That is as it should be, but it slowed our project down considerably. The people who are using the product like it very much. It was rolled out two weeks ago. We have all of the transaction information on 281,000 of their claims in our system and several million medical bills. Our concern is there is new management in this area of the company. Until they settle in, it makes us nervous.
However, I am absolutely convinced that by next summer the development of this product would be to such an extent that people will not be able to stay in the business without it.
Krish: So, things are proceeding well with Marriott then?
Bob: Yes, we have delivered more than promised and the system is performing very well, actually exceeding our expectations. In spite of the management changes we believe are in for the long haul with us. We also have a bunch of other potential customers. We have the enterprise sales problem. One customer is in middle of changing their claims system and we have to wait 18 months before they can make a decision. Others will be making decisions in the next couple of months.
Krish: Have you gone out and looked at opportunities broadly?
Bob: We are talking to other Fortune 500 companies. One company was changing their TPAs last year and asked us to contact after a year. They asked for a contract to do a paid pilot with their claim data. They are a larger company than Marriott. We are also talking to group self-insurance pools. One such pool is handling claims for 34 school districts in California. They pride themselves in being an early adopter so they could lead to a great deal more business among California school districts.
Krish: If you had to rethink your strategy for Go to Market, what would you do different compared to a year ago?
Bob: Well, I don’t think that we would have changed our GTM strategy. I would have probably raised a lot more money early. That would have helped put a couple of extra developers on the team. The product will sell itself once we get past a critical set of features. Doc Compare for example. It could be extended to compare anything. A customer in California can’t believe that something like Doc Compare is possible, but we have built 80% of its features already.
Krish: Isn’t running a product dev team different from managing technical projects in your previous experience, such as actuarial services?
Bob: In the subject matter it was different, but in basic function it was the same. You have to define a goal you want to get to in a sufficiently specific way, so people know where they are going, and not so tight that you foreclose additional learning and creativity. So, you may want to pivot a little bit and emphasize one feature over other. What I enjoy most is getting really talented hardworking and honest people in a team focused on a goal. That is a lot of work and worry, but ultimately it is the most fun.
Also, we are not really worried about competitors entering the market. Competition validates the idea of Claim-Maps. With the team we have got, we think we can run faster than anybody else.
Krish: Is data formats from different customers an issue?
Bob: Fortunately, we don’t have as many problems as others do in data integration. It is not without its challenges. For the most part, the data coming from legacy systems is surprisingly clean. Consistency with data definition is also good. Most are 837 transactions, which is a standard. We also get financial information after the bills have been paid and that is very clean as well.
We do have some challenges with the data, but it is nowhere the magnitude the people outside the industry expect.
With Marriott, we even proved to them that they had paid before the claim was filed. This experience also proved that Claim-Maps can show you things that you thought was not possible in your own systems.
Krish: So can one expect consistency of data between say a Walmart and Marriott?
Bob: yes it is surprisingly consistent
Krish: Bob, What would be your advice for entrepreneurs?
Bob: Here is what I say 1) Be tolerant of uncertainty. 2) Keep selling all the time even starting from when you just have a PowerPoint. For enterprise sales, talk to a lot of people so that you are in their queue. The queue could be six to twenty four months long. 3) Raise as much money as you can so you can over invest in your product to get past objections.
I would cite an example of Jeff Bezos who applied for Amazon to be granted self-insured status. To be self-insured the state requires a net worth of at least $5m, but at that time Amazon was -$1b. That was a classic example of over investing that made the uncertainty a lot more tolerable. If Jeff had not raised as much money as he did as early as he did, Amazon would have never survived long enough to take over the world of retail.
Krish: What are your goals for 2013?
Bob: Our initial release is done. We want to close 2 more customers and it does not matter as to their size. We need a minimum of 5 customers to be interesting in this space to make raising money to be not a problem. We will need less cash than most companies because our customers are used to paying up front in advance of service delivery. This is insurance and everybody pays in advance in this business.
We are in the middle of an extended Series A right now. We are confident that several of our preferred investors will over subscribe. We may raise this in a tiered pricing structure with the first $400k at a more favorable price than the remaining $600k.
Krish: Bob, it has been a pleasure talking to you. Wish you and Claim Maps all the best in the coming years.
Bob: Thanks. I enjoyed it too. We have some great stuff coming in the roadmap.