Recently I wrote in a blog post about a startup called Aviva. My comments were based on an article I read. According to the CEO, “What’s our long-term goal? To go from Ask it Once to Ask it Never – so customers don’t have to answer any questions at all.” How can coverage be booked without asking ANY questions? Why, using big data, of course.
Wouldn’t a better goal be to ask the necessary questions to assist consumers in identifying their unique exposures to loss, then match those exposures where possible with the proper insurance package to minimize the likelihood that they will experience a serious or catastrophic financial loss?
At my semi-annual checkups, my doctor asks me a LOT of questions. Would it be an improvement if he didn’t ask me ANY questions? Maybe for his bottom line, but not for mine. Who can’t spare an hour once a year to prevent financial ruin?
In another blog post, I wrote about the startup, Slice, which apparently plans to write on-demand home sharing and ride sharing insurance without an application. How? Presumably using big data, of course. In still another blog post, I wrote about Lemonade which writes homeowners insurance using a phone app without a lot of pesky questions designed to identify exposure gaps of individuals and families. They too seem to be relying on black box algorithms and our friend “big data.”
Let’s take Slice for example. They claim:
“All the information that insurance carriers ask you is all publicly available. So instead of taking up your time to give us this info, we use our clever SliceBots to collect it.”
So, ALL of the information they need to properly insure all of your unique exposures to loss is publicly available? At one time, I saw a Zillow logo on one start-up’s web site. Is that where, for example, homeowners information might be obtained? Or might such a startup go directly to tax and other records where this information is obtained? How reliable is this “big data”? Is it vetted out all if customers are not asked any questions?
Still another startup, as discussed in this Forbes article, is Hippo. Backed by a number of investors, including Trulia, this is how their big data approach works:
“According to the company, with Hippo, consumers can go from quote to purchase in minutes, as quotes are delivered in 60 seconds after answering three simple questions. Customers can get a personalized Hippo quote online, by phone or even through Facebook Messenger. The company leverages technology and data from multiple sources (such as property records, permit filings and aerial photography of roof conditions) to streamline the application process and provide ongoing risk monitoring. By leveraging data, Hippo saves customers time, while also garnering more accurate information that cannot be provided from subjective human answers alone. By cutting out the middleman, more accurately assessing risk and increasing technology efficiencies, Hippo is able to pass savings on to consumers.”
There happens to be a home for sale in my neighborhood. Out of curiosity, I checked it out at Zillow and Trulia. Zillow says it’s a 1-story home, Trulia says 2 stories. Zillow says 2 ½ baths, Trulia says 3 ¼ baths. Zillow says the lot is 1.6 acres, Trulia says 0.48 acres. Zillow says the home is 2,968 sq. ft., Trulia says it’s 3,891 sq. ft. Just in the replacement cost valuation of the home alone, think these discrepancies might make a difference in coverage limits?
In my own case, I owned a home that was 1,000 sq. ft. larger than the country tax records showed. Over the course of 30+ years, attic space had been converted to living space, but the records from which “big data” might be drawn were never updated. When discussing this issue in an online forum, one of the participants said that Zillow showed his home being 2,400 sq. ft. (the same size in the tax rolls), whereas it’s actually 4,683 sq. ft.
Big data is one thing. Big, BAD data is another. Who is vetting this information, bots and algorithms? Certainly not regulators given the open-arms welcome one startup got from a state insurance department. Is anyone listening? Does anyone care?