Here's something you may not know, Henry Ford built neither the first cars nor the best cars, he did not invent the internal combustion engine nor any of the parts that comprised the first automobiles yet he was able to dominate the auto industry. How was that possible? Well, Henry Ford was a genius but, this is where it gets interesting, his genius wasn't for cars, it was for process. His obsession and brilliance for process led to the development of the automated assembly line which enabled him to build product faster and cheaper and that's how he bested his rivals. Ford proved that process can sometimes supplant product as the winning factor in a competitive industry.
Of course, in the Insurance Industry any factoid about Henry Ford will be purely anecdotal but there are important lessons to be learned here. Over the past 15 years advances in enterprise platforms, digital applications and big data technology made it possible for insureds, using web based applications, to receive a policy in minutes after entering a few bits of data. In some cases it was possible to scale back or completely eliminate the need for middlemen such as brokers and underwriters. This was a technical revolution which the insurance industry labeled Insurtech. It was a revolution that brought with it the promise of better products and greater efficiency. But like most revolutions, things invariably went sideways. Many of the pioneers of Insurtech platforms focused almost exclusively on user interfaces and front end systems, this is what impressed the VC investments that were pouring into the industry. Back end processes such as policy administration, premium accounting and claims processing were largely ignored. Then the failures began, Willis Re in their 2021 quarterly update noted that there were 456 Insurtech failures over the past decade and high on the list of reasons for these were operational failures. This is the equivalent of those early car companies that failed because they were building great cars but did not have the solid processes in place to support them.
But if an Insurtech company has a great product, shouldn't they automatically succeed? Well, not necessarily, the industry has gotten to the point where it's no longer a question of who has the best product or the most innovative platform. It's now a question of who has the best underlying process, and ultimately, it's a question of simple math, something that insurance companies are supposed to be good at. The mathematical reality facing Insurtechs is this, if you're processing thousands of transactions per day you cannot afford any failure in your back-end systems. Take for example an Insurtech company, call it Company A, a high volume, low premium, Auto insurer which processes 30,000 transactions per month. An error in any of the back-end processes such as policy issuance, premium accounting or claims processing means that 30,000 transactions have to be unwound, for most companies that's manageable. Let's say the error goes unresolved for 3 months - now you're dealing with 90,000 transactions not ideal, but still manageable. If, however, the problem persists for 6 months maybe a year now you're dealing with hundreds of thousands of transactions that have to be unwound and corrected. This is a huge undertaking often requiring outside consulting help, system upgrades, additions to staff etc. the costs here are tremendous and that's before you add in the loss of business due to angry customers. For many of the Insurtech failures citing operational issues as the main reason for folding this is what they mean. The unanticipated costs of fixing broken processes had a huge impact on their bottom line, an impact that many of them were unable to recover from.
If broken processes have caused a large number of failures within the industry then why are they still ignored? One of the reasons is that there have been no high-profile failures, no single headline grabbing incident that can serve as a cautionary tale to the rest of the industry. For that we have to look at another equivalent industry that is similarly process driven, that is digital and uses large data sets and that industry is Crypto. Here we do have the headlines and probing questions and this is where the caution to Insurtechs lie. In an interview with Ross Sorokin of the New York Times Sam Bankman Fried reflected on the failure of FTX and commented that:
" We started out by trying to out-build our competitors, what I realize now is that we should have tried to out-process them"
Now, before I get a barrage of hate mail, please note that I'm not in the habit of quoting alleged fraudsters but there is something about this statement that rung true for me. Crypto like Insurtech has attracted investment capital by focusing on models, algorithms and front-end systems. There was little regard for the back-end processes that was supposed to run these or the cost if these processes were to fail. SBF thought of it but he was a day late and 8 billion dollars short. This is not to suggest that Insurtech will suffer a similar fate, it won't, largely because it's supported by deliverables that are more substantive. But the lesson here is that FTX's moment of truth underscores an important point not only to Crypto but to Insurtech as well - that focus on process can not only avert disaster but can perhaps become a winning strategy.
There is another reason why processes are ignored and it has less to do with headlines and more to do with mentality. Insurtechs are somewhat crippled by their perception of the industry that they operate in. The Insurance industry has historically been framed as plain and boring, a place where numbers nerds pour over actuarial data. Insurtech changed that, they were perceived as cutting edge and cool. If you don't believe that then look at the insurance commercials on television, they are quirky, fun and savvy. And this is the image that brings in customers and investors. As a result of that perception the vast amount of investment in Insurtech went to the revenue generating algorithms and front end systems with little left over for back office processes. Insurance companies often have pre-defined ratios of how much capital they would like to invest in revenue generators vs back end processes these ratios range from 60/40 to 70/30 to 80/20. But here's the rub, Insurtechs have a processing component and the lower the operations part of that ratio becomes the less you operate as a true Insurtech. With little or no back office processing Insurtechs essentially become software companies with a cutting edge front end supported by an algorithm for pricing risk. Now, this is not necessarily a bad thing but keep in mind that software companies have a short shelf life (about 5 years) then you need an exit strategy. This identity crisis of thinking you have an Insurtech when you have a software company often leads to failure because the company is being viewed as one thing but managed as something else.
Even in cases where Insurtechs are attentive to back office processes there are gaps because of the tendency to focus on the obvious. The processes that are broken or are prone to being broken tend to get all of the attention. However there are often hidden, underlying issues that need to be addressed but due to the inconspicuous nature of these it's difficult to explain why these need to be fixed. How do you even frame a conversation about the need to fix problems that are not obvious to decision makers. Interestingly enough a possible answer to this comes from an unlikely source - The Smithsonian Museum. Stick with me on this, there is a point here I promise. Buried among the thousands of exhibits at the Smithsonian National Air and Space Museum in Washington DC is the following illustration with a compelling the story behind it.
Recommended by LinkedIn
During World War 2 a young mathematician named Abraham Wald was given this diagram with a note " these are bullet hole patterns on B-25 bombers that survived raids over Germany" Wald was then asked to come up with recommendations to reinforce those areas. What Wald proposed instead was reinforcement of the areas without bullet holes. His reasoning, planes that were shot in the areas without bullet holes did not survive, so they did not return for the Army Air Force to examine their bullet hole patterns, if those areas were reinforced then those planes will have a better chance of returning. This reasoning was later developed into a theory called survivorship bias.
Insurtechs are not immune to survivorship bias especially when it comes to systems that are fairly expensive to maintain, there is an organizational focus on getting these systems to produce what they are intended to produce and because of this other processes and systems are often overlooked or are subject to the mentality "if it ain't broke why fix it?" Getting corporate buy-in for an organization to focus on the seemingly neutral areas of their process (the area of the airplane without bullet holes) requires a serious discussion on something that goes along with process and that is controls. If you look at problems the way that Albert Wald did then having a conversation about processes without a follow-up on controls seem non-sequitur. Its the logical follow-up to building out a a resilient, pro-active back office process. Establishing controls around process can help to identify problems in areas that are overlooked.
Now, it won't be fair to write an article on process deficiencies and failures and not offer any solutions. Over the past 10 years I've been involved in the financial operations implementation for 33 MGUs and Insurtech startups. And if I were to take all of my observations and boil them down to a single diagram of what's needed to avoid process failure it will be this
This is the barebones essential for running a successful back office process for Insurtechs. Typically the first question that I seek to answer when I work with a new company is how much of this process flow has been implemented is it 60%? 70%? 90%? As a rule of thumb if these processes have not been at least 90% implemented by the first year of operation then the chance of failure increases exponentially. To perhaps overstate the obvious, this process flow is built on the understanding that premium is the lifeblood of any insurance organization and success or failure ultimately depends on how well premium is managed. Timely issuance of borderaux and premium payments are necessary to maintain the trust of your carrier partners and timely commission sweeps help to minimize the cost of operating capital. All of this is dependent on timely and accurate reporting but the systems needed to generate this reporting are, admittedly, not cheap. Many underwriting systems with an attached accounting module can cost upwards of $30,000 per month and that excludes the cost of personnel to run it. Or, If a company has the inhouse technical talent, implementing a cloud based, business intelligence product like Amazon Web Services QuickSight with an underlying data warehouse, can be a cheaper alternative. Either way costs are unavoidable, but instead of focusing entirely on costs perhaps a better question is -can an organization afford not implement these processes? Is the cost savings worth the loss of trust with carriers and reinsurers due to delayed or inaccurate reporting, is it worth the loss of investment income due to delayed commission sweeps. The answer to this largely depends on the organizational goals and resources of a particular company. But regardless of what those goals are this certainly something that needs to be considered.
Controls are also very important. Processes without controls are often ineffective. In Insurtech, controls can be something as simple as closing an accounting period or ensuring that invoice and check numbers are not duplicated. Failure to do these two simple things can result in duplicate payments, financial statement misstatements or missing information in the data provided to carriers/reinsurers. Of course controls are not relegated to just the simple ones cited above, there are many more considerations than these. In general, there are two necessary types of controls, detective controls to catch errors and preventive controls to ensure that those errors don't happen in the first place. Controls can, admittedly, be clunky, tough and expensive to implement but controls, in all of its tedious forms, can also save a company from financial ruin. Every Insurtech company should have serious discussions about controls and invest the time and money to implement these.
The third factor in implementing an effective back office process has less to do with systems and more to do with personnel. Insurtech is replete with people offering financial operations solutions but many of these are from other industries and don't seem to understand the nuances of insurance operations. What's worse is that they have no clue that they have no clue and forge ahead with solutions that don't fit the insurance model. The main point to keep in mind is that, ours is a particular industry so it's important to seek out the people with the right experience because they will be setting up processes and systems that your business will have to live with for the foreseeable future.
So, processes, controls and having the right back office personnel are the three main factors in ensuring that Insurtechs address a pressing need in the industry. Of course there are other necessities and back-end processing isn't the only reason why Insurtechs fail. My friends in the industry will point to other reasons for failure such as under-capitalization, poor timing and poorly executed strategies. All of these are valid factors and are worth considering when discussing Insurtech failures but the main reason cited among failed Insurtechs is operational failure and that's why process and operations were the main focus of this article. Thank You.
About the Author:
Manooj Persad is the founder of JMAD Consulting which is an advisory service providing financial operations solutions to Insurtech, MGA/MGUs and Insurance startups.