In P&C insurance, location has a profound impact on risk. This is not new news. Insurers have been using location as a factor in risk assessment and policy pricing from the inception of the business. A house in the middle of a meadow has a much different risk profile from a house on the beachfront — unless, of course, the meadow is located in tornado alley.
The terabytes of data available today to insurers (and InsureTechs) can allow for a much more refined assessment of risk — but that takes precision data that is easily analyzed. The concept I want to share is location genealogy — a location intelligence model that can help organizations programmatically uncover the relationship between addresses, locations and buildings.
Consider a simple example: a sandwich shop in a low-crime area that is not prone to flooding — risk factors that you can often determine just by zip code. It looks like a profitable opportunity. But what if the shop were in the same building as a fireworks store? Would that impact your pricing? Would you still write the policy? Would you even know?
Defining location genealogy
Without a site inspection or taking the time to call up the Google street view, probably not. And that’s where location genealogy comes in. It takes advantage of the three pillars of Location Master Data Management to programmatically link properties at the building level to reveal relationships you might not otherwise know exist.
The three pillars are a valid, consistently formatted street address; a hyper-accurate geocode; and a unique identifier permanently and unambiguously associated with the place, even if the address were to change. We call that identifier the pbKey™, and we have one for every address in the United States. We also have a unique identifier for every building. We have associated every address identifier (the parent, in terms of data structure) with its building identifier (the child).
This data structure makes it easy for underwriters, actuaries and data scientists to use existing software tools or machine learning platforms to identify and understand these relationships and gain full transparency into a property. Here are three use cases for location genealogy:
Identify different addresses located in the same building
It’s not unusual for a multi-tenant building to span an entire block, or half a block. That means the businesses or people housed there could have addresses on three or even four different streets. Without location genealogy, it would be difficult to know that those addresses actually belong to one building, making it difficult if not impossible to accurately evaluate the consolidated risk existing under one roof.
Evaluate your current book of business
Using pbKey and building ID data, you can build a property database of your current book of business. Then use location to evaluate if you have more risk exposure in an area or a building than you want. For example, five insureds under one roof could be unacceptable risk in the case of a building-wide fire. Conversely, you can use the data to identify prospects located near existing insureds or in areas where you have no coverage. Know in advance who you are marketing to, and what the risk profile is, before you invest in contacting them.
Consolidate risk data in one file for efficient processing
With Location Master Data Management, you can consolidate a broad range of data attributes that affect risk, all of which are linked to the pbKey, into one file. That means for any one address, you can drill down and create a thorough and highly accurate risk profile without having to spend extensive resources on data preparation. With this level of precision, you can identify the risks you don’t want to take on — and you can identify those potentially profitable low-risk opportunities located in high-risk areas.
To learn more about how Location Master Data Management and location genealogy can help insurance underwriters, actuaries and data scientists understand exposures, reduce risk and increase pricing efficiency, read Mastering Location Data: Close, But Not Quite There from Harvard Business Review Analytic Services.