The data edge in insurance

I’ve had the pleasure of being in a few different industries now: call centers, industrial machinery, and now insurance. It’s always interesting to try to observe not just the differences between them but also their similarities. Specifically, the similarities in how these different industries are incorporating data into their vertical stacks.

Within insurance, data has probably a more direct impact than any other I’ve seen because it’s based almost entirely on math, at the end of the day. You have to take in more than you pay out.

Here are a few areas where there is just so much potential for data in the insurance industry, and some interesting things different people are doing with it.

Using data to underwrite better

Boils down to taking more good risks than bad

  • classic data science application, write algorithms that are smarter
  • finding completely new sources to use such as drones, sensor data, and social media profiles
  • partnering with third parties who have exclusive access to useful data including personnel histories, vehicle location, business profitability, and more

Ultimately all of those come

sources of proprietary loss data


  • Lack of a process for figuring out how to use new data sets
  • Getting new data sets into a data lake

Using data to simplify insurance for the consumer

Prefill data etc

Using data to simplify insurance for the consumer


Data Companies Eat Software Companies

I was in the car this morning listening to a BBC report that, for the first time in a long time, the mortality rate in the US went up instead of down. The average life expectancy actually got shorter this year. That’s scary, but it got me thinking about software and a trend that I’ve noticed. A trend that is going to separate the winners from the losers over the next five years.

There’s a clear line developing between the future haves and have-nots in the digital world. A line of thinking around where the value lies and what you can and cannot make money on.

Traditional software companies have focused almost exclusively on building software and charging for it. This is outdated thinking and looks to me a lot like painting a giant target on your forehead. Using this model today is like charting your course by looking in the rearview mirror. Meanwhile you don’t realize that you’re going off the road.

If you look at what the most forward-thinking companies are doing you can start to see glimpses of where the world is heading. Here’s a hint: they’re all giving their software away. Why? Because software doesn’t matter.

Have you been watching as tech giants who understand where things are going give away all of their best technology? Look at Google, who isn’t locking its AI technology, Tensorflow, away in a deep, dark vault — instead it’s giving it away. Same with Facebook, who gave away its React technology and looks to be headed in the same direction with its own AI stack, FBLearner.

Why? Because on a long enough timeline all technology becomes commoditized, and they know it.

The reality is that all of the supposedly “hard tech” is becoming more accessible by the day. Most if not all predictive algorithms are out there for anyone to use. Scalable architecture is getting easier to build every day. There’s really nothing you can’t do with open source technology now, the barriers to entry are dropping constantly. Some of the technology that’s out there right now for free just blows my mind. Tools like OpenAI Gym would have been sold for megabucks just two years ago. They’re free now.

So what does matter then? Where’s the real action?

Control over the data. That’s what matters. You can replace technology, you can’t replace data you don’t have.

The kingdoms of tomorrow will be built on an empire of data, and the land rush is happening in plain sight, right now, if you know where to look.

Most companies won’t come out and say that data is the most important thing, but actions speak louder than words. When IBM acquired and Merge Healthcare for multiple billions of dollars, many people were puzzled. There’s no obvious synergy there, the acquisitions happened because the value was in the data.

The most interesting companies today are working hard on giving away software that other companies charge tons of money for. Why? In order to get access to the data.

I agree that software is eating the world. But in the technology food chain, data eats software.

The really interesting dynamic here is that all of the software-driven startups are effectively creating a hit-list of areas where free technology can come in and take them out at the knees, undercutting them in order to get the data. I suspect we’ll see quite a bit of disruption to the status quo in the coming years, there are lots of companies out there who aren’t thinking about this correctly.