How to build failures (and get away with it in the insurance world)
I have not failed. I’ve just found 10,000 ways that won’t work. (Thomas Edison)
Introduction
It is estimated that there are about 200 millions active companies in the world. 45'508 companies are listed in the stock exchanges. Fortune Global 500 lists — well — the globally 500 largest companies in the world. Currently, 4 companies are valuated more than 1 trillion dollars. Obviously, the larger and more important a company, the small the number of companies you will find. But the relationship is not linear. If you group the companies with similar valuation and plot the size of each group on a graph, you get the following pattern:
In other words, you have many companies that have a low valuation and a few companies with extreme high valuation.
Why is it interesting? The beauty and the power of this curve is that it is a recurrent pattern, called the power law. Once you learn to see it, it’s easy to spot it everywhere:
- frequency of words
- magnitude of earthquakes
- retweets on Twitter
- size of cities
- commercial success of insurance products
The latter case is of particular interest to us. Measuring the commercial success of insurance products can be achieved in different ways. For the sake of simplicity, we can say that an insurance product is successful, if it has sold a many policies over a long enough period of time. If you ask any large or small insurance and reinsurance company, it is easy to validate the claim that the power-law applies: most products have disappointing sales volume, while a few spread in the market and become a hit.
Embrace the power of math (if you see it)
Actuaries deal with the study of risk. The first and most important tool at their disposal is the law of large numbers. It is the law of large numbers that allows insurance companies to take risk and to exist at all. It is a very well known theorem in probability theory the underlying probability rate manifests itself as the average when an experiment is repeated many times. To put it simply, you can estimate with high precision the amount of claims you are likely to see in a certain portfolio, if the latter is large enough (and other preconditions apply). Computing a corresponding premium to sustain the outflow and your costs is then a piece of cake. The faith in the predicting ability of this law is paramount to actuaries and insurance professionals and has never been questioned. Discussions arise only to question if the preconditions of the theorem are met in the portfolio at hand (read: antiselection, trends, …).
In the insurance company we run, we decided that ignoring the power-law nature of the commercial success of insurance products is as foolish as to ignore the law of large numbers. Would you believe me if I were to tell you that I rolled a fair dice 1'000 times and got an average of 5.5 or above?
At Squarelife, we decided to play the following thought-game: as a decision-maker, product developer, actuary, … how would you behave if you would know (within a reasonable likelihood) that the next product you are developing is going to flop in the market? Bear in mind that this game is not artificial, it grounds in the same principles as the law of large numbers! It is as likely as to know that the dice above would average to around 3.5.
The only option is to fail (many times)
When you share this theory with professionals in the insurance world (which is of course not the only industry affected by these ideas), you might have the strange feeling you are back in the Middle Ages, trying to explain to others that the earth revolves around the sun. At best, it looks like a funny anecdote to make jokes about.
Once you realize and believe that you are fighting with all your resources against a natural law, you might feel enlightened, and you can start embracing it. Since the frequency of good products will follow the pattern of the power-law, the consequences are two-fold:
- the next product you are working on right now is likely to be a failure
- if you keep launching products, you will eventually hit the long-tail of the curve and have a success
This realization has been an eye-opener for us. It all boils down to “survive the next product” and wait for the success to come. Easy, right?
Skeptical people usually argue that if a product fails, you might have done something wrong, most likely bad marketing (bad communication, wrong channels, not enough money). We are not claiming that the success of a product is a pure event of luck, of course not. Prepared professionals, good marketing, the ability to know and listen to your clients are all elements that can change the slope of the power law curve, increase the long tail and ensure you a better return on your investments (or not). Note that we have just mentioned that you can change the slope of the curve, not its form!
Learn how to fail
It might be a rather pessimistic view of the work we do, but it is not. The key is to swim with the flow dictated by math and not against it. You should focus on a strategy that allows you to sustain failures instead of blindly betting for a good outcome on the product you have been developing over the last two years.
We think that there are many ways to be efficient at this game and that we could find a few principles and good practices that work for us. More on this topic in a next article.
I would like to thank many friends for their precious inputs on this artic
-Elias, Squarelife CEO
