The news this week that Amazon is looking into some form of home insurance product seems fairly speculative, but it’s not an insane idea. The Information reported the story, with a lot of disclaimers, and quoted one anonymous source at Amazon who said the company has had discussions about selling insurance as part of its smart home devices business.
For years, the insurance business has seemed like it might become a big channel or driver for connected devices. Insurance companies have partnered with firms such as Nest, Canary, and Ring to offer discounts to clients who put such connected devices in their homes, for example. And two years ago I spoke with a number of insurance companies that wanted to send policyholders home with a box of connected gear that could help decrease the risk of fire, floods, or break-ins.
The cost of connected home gear, and the confusing standards that make it such an intimidating prospect for mainstream consumers, seemed like an opportunity for the insurance business. The thinking was that insurers could draw a clear line between putting a water sensor or a smoke detector in a home and lowered claim payouts. With that insight insurers would subsidize the cost of connected devices in policyholders’ homes, maybe even send them a package of gear when policyholders signed up.
While a consumer might balk at paying $600 to install six connected smoke detectors throughout their home vs. paying $90 for a set of “dumb” sensors, an insurance company would save so much in potential claims that the $600 in smoke detectors would be worth it because the insurance company could confirm that they were working or not.
So far, after a rush of trials in 2014 and 2015, that hasn’t happened. Insurers haven’t really moved beyond testing connected devices in exchange for a discount on premiums.
Brett Jurgens, the CEO of Notion, which makes a multi-purpose connected sensor, blames it on a lack of data. “Insurance firms know what factors make for a safer driver, but we don’t yet know what makes for a safer home.” He says that insurance industry watchers were excited about the potential for smart devices to change the risk profile for insured homes, but that is still far off.
The problem is that insurance firms are run by actuaries, says Roel Peeters, CEO of Roost. And actuaries focus on risk. “What is ingrained in the insurance company culture is that you have to be able to price risk,” he says. “You must be able to say that the impact of this risk is xyz plus or minus two cents.”
But a company first has to invest in the tools to get that data. And smart home products are expensive, which means that setting up those trials can cost millions. For example, an insurance firm needs 50,000 years of data on water leak sensors to determine if they can help reduce risk and by how much. But that kind of actuarially relevant data means a company has to put 50,000 water leak sensors in 50,000 homes for one year (or put 1,000 sensors in 1,000 homes and wait 50 years).
Peeters says that for his products that would cost roughly $5 million, which is too much for a startup to invest and even too much for insurance firms. “That is a significant business decision,” says Peeters. “If I have the data, I can easily justify spending this money; if I don’t have the data, I can’t make the business case.”
Roost is working with a cadre of insurance firms to try to aggregate data, which they would gather by splitting the 50,000 sensors and homes among themselves. From there, they hope to derive the data they need and then offer products that take advantage of connected devices, but in ways that go beyond one-off discounts.
In the meantime, Peeters sells Roost devices to insurers as a means to differentiate their policies. Jurgens from Notion is doing the same thing. A consumer has to love the device and find daily value with it, however. With a water sensor or a smoke alarm, that might be tough, whereas security-related products have an edge, which is why Roost is coming out with a connected sensor that can tell homeowners if their garage door is open or closed.
Both Jurgens and Peeters seem bemused by how archaic the insurance industry is. Most insurance providers have little insight into their customers, and in many cases only know their names and home addresses. That’s because the end consumer buys a policy from an agent, not the insurer. So for many insurance companies, even something as simple as being able to detect that an alarm system the insurer has given a discount on, has been turned would be helpful.
Even that can be a challenge, because for many insurance companies, getting that data relies on being able to tap into a huge number of different devices and systems. ADT and Nest both have connected alarm systems, each with different features and offerings. Trying to get the status on those things requires customer trust (they have to opt-in to sharing it with an insurance company) but also technical skills to deliver that information in a meaningful way, says Patti Griffin, chief product officer with Duck Creek Technologies, which provides software to the insurance industry.
Trying to use data points from disparate systems is also a challenge, which means that when we finally see insurance offerings they may only exist as a limited and tested number of connected devices. But it will be a start.
So, while the insurance industry lumbers into the modern era, it appears we’ll have to wait for it to pick up the tab for smart home devices.