Sapiens TL
Follow Us



Accuracy: Are Data Providers Pulling the Wool Over Your Eyes?

Eye of the Storm: Natural disasters, the insurtech market, and other musings from HazardHub

For the longest time, the insurance industry has dealt with, “close enough is good enough.” And the law of large numbers is partially to blame. It’s a good law, and it works. But it covers up a lot of poor decisions, or decisions that could be better, given more accurate data. 

This is a very simple concept: How you look at accuracy determines the way you build your data. No one ever builds data to be inaccurate. We’ve been building geospatial hazard data since pretty close to the creation of geocoding. Because while it was nice to have dots on a map, it was more important to understand what was around those dots. Geocoding has moved from ZIP codes to census designations, to interpolated street sides, to parcel centroids, to building centroids, to corners of buildings, to soon adding in the Z coordinate to tell us how high. This is a simple progression to greater accuracy. And as the precision of geocoding has gone up, overall, the cost of geocoding has come down. 

Most people’s approaches to data are directly influenced by their data background. No data are perfect. But people in professions where perfection is the goal are going to produce data in a different manner than those that aren’t. One overwhelming belief that a lot of people have is that more precise data costs more. It’s caused many an internal price/ROI discussion on the value of better data. For the record, better data are invaluable, but it doesn’t have to cost more. Sometimes not using better data leads to higher costs.

Let me provide an example: fire hydrants. When signing up for a property policy, everyone wants to know whether or not there is a hydrant near the property and how far away it is. For some people, simply saying that the property is inside a municipality, and therefore has hydrants, is good enough. Having built the only national hydrant database, we can tell you this assumption is way wrong. But it doesn’t cost much to make that determination and you’re certainly getting what you’re not paying for. 

Others can provide the number of hydrants in an area but not their location. This costs more and assumes an even distribution of hydrants throughout the community. Again, when you’ve built the real data, you know how wrong that assumption is. Interestingly, knowing the distance to the nearest hydrant and the number of hydrants within a 1,000-foot radius barely costs more than the hydrant assumption of being in a municipality -- and costs significantly less than just having the community hydrant count.

Data providers to the insurance industry have pulled the wool over the industry's eyes for a long time -- they rely on the law of large numbers as much as insurers do. Even catastrophe models, thought to be the paragon of accuracy, take exceptionally high-level data sets and try to force them down to the property level. It's as ridiculous as it is wrong.

Building accurate data from the bottom-up results in cleaner, more accurate results.  And now, thanks to better inputs and construction methodologies, they’re significantly more cost effective.

Featured articles

Guidewire MR

Clear Lync MR


The Email Chat is a regular feature of the ITA Pro magazine and website. We send a series of questions to an insurance IT leader in search of thought-provoking responses on important issues facing the insurance industry.


April 5th – 7th, 2020
The Diplomat Resort
Hollywood, FL
Become a member today to receive updates –


only online

Only Online Archive

ITA Pro Buyers' Guide

Vendor Views

Partner News