Removing noise from small commercial insurance data
Bad data is worse than no data. Review how to cut through to the useful insights insurance data carries.
While the volume of data available today can vastly improve a range of insurance processes, the tidal wave of information can be overwhelming. This is particularly when trying to figure out what is useful and what is just noise.
According to Verisk Analytics, Inc., this ability to cut through that noise and get to the most useful insights is particularly vital in small business commercial lines, where customers and agents are seeking quick quoting processes.
In fact, nearly 40% of agents said they would take a lower commission rate in return for a seamless transaction with an insurer, “especially if responses to the application were dependably accurate and delivered quickly to the insurer,” according to Verisk.
“Bad data is worse than no data,” says Tracey Waller, director of small commercial underwriting at Verisk, telling PropertyCasualty360.com it’s not only important to work with a variety of datasets, but those of the highest quality as well.
For example, Verisk works with crowd-sourced review site Yelp to get a better picture of the true risk profiles of small businesses. Waller explains partnering with a site such as Yelp can vastly enhance the available data because social media is where small businesses tend to deploy marketing. In turn, the data, including photos, is current, complete and correct.
“Finding current and robust information online to underwrite small commercial insurance can take significant time and effort,” Waller said in a release announcing the partnership. “By working directly with Yelp, Verisk is augmenting its high-quality and consistent analytics on millions of small businesses with information that is up-to-date, organized and easy to digest.”
Take a plumbing company for example. Perhaps it started out with draining cleaning and repair services, but grew to include more intensive work involving excavation or other work that would change the risk profile. While the business owner might not have thought to update their insurance provider, they likely advertised those new services.
To unearth better data, Waller says it is critical to toss out of date & bad data. Incoming data can be incomplete or out of date, making validation a vital step as poor data sources can result in false impressions of a business’ risks.
“Validation should be ongoing and never-ending,” Waller says. Even after drilling down into a dataset, irrelevant information can remain and should be filtered out.
Related: