In our continuing “Connected” series, Verisk IoT experts Dawn Mortimer and Drew Doleski discuss best practices for launching a smart home pilot initiative.
Before you start: some important ‘Dos and Don’ts’
A smart pilot can be a starting point for carriers who want to test hypotheses about the role of the smart home in insurance and to better understand the potential for leveraging data. Verisk’s Dawn Mortimer and Drew Doleski have a unique perspective on some definite “Dos and Don’ts” to consider before launching a pilot. Prior to joining Verisk, Dawn led numerous smart home pilots with insurance carriers and Drew was responsible for insurance industry partnership programs at a global smart home brand.
So what are some of the "dos and don’ts"?
Do identify key learning objectives
Drew: Dawn, I think it’s hard for many insurers to get a full understanding and appreciation for the kind of knowledge they can glean until they do a pilot. Looking at your pain points and growth opportunities is always a good place to start. Are you seeing lots of claims from water, fire, intrusion—from any of those perils that Internet of Things (IoT) solutions or alarm companies can help with? Is there a new market segment that your underwriters would like to unlock that might be enabled if you had visibility into the risks using IoT data?
Dawn: I agree: Look at your pain points. At my former company, water was the number one peril we were focusing on, and today in talking with a lot of carriers, it seems to be their number one peril as well. They’re seeing frequency continuing to rise and severity isn’t lessening. Last year, Verisk’s Xactware division reported 90,909 non-weather, water-related events. The average water claim was $7,734, and we're seeing, on average, 2.5-day lag times from the time an event occurs to the time it gets reported to the insurer.
Do focus on internal partnerships
Drew: Those are interesting stats. You also point to the need for tight alignment between whoever is managing the pilot and the business units who stand to benefit from the results or will become accountable for operationalizing the program. Once you’ve figured out who needs a seat at the table, it’s important to collaborate with them to articulate the learning objectives that will be most beneficial. The best ones are always measurable!
Dawn: Absolutely. In my experience, it seemed we were most successful when I went to the business—especially claims—and asked them to name the biggest challenge they saw. As I mentioned, they told me water claims were rising more than ever. I recommended that if you want early detection of water risk, start putting data collection kits in customers’ homes. Perhaps start with people who’ve had past claims, because they know the pain of experiencing a water event and will probably be more amenable to installing a kit.
Drew: Once you understand the problems, partner extensively with the business unit leaders across your organization so you can quantify what the issue is, and cross-functionally design a pilot where you know on the back-end that you’re going to have sufficient volume to generate actionable insights.
Dawn: You also then have a business unit that will be more likely to assume ownership beyond the pilot. You can run as many experiments as you want, but if you can’t operationalize your programs and make them part of the everyday workflow for the underwriters or adjusters, you’re not going to be successful.
Drew: And don’t forget to include your front-line folks too! You’d be surprised how many customers will reach out to Customer Support or their agent when they have a question about the technology that you provided them—giving these internal teams a heads up about where to route the inquiry is always helpful.
Don’t underestimate data needs and pilot scope
Drew: Once you understand your learning objectives, you can start to outline your experiment. I recommend reverse-engineering the scale, data volume, and geographies that you’ll need to be successful.
Dawn: Good point. When I talked to people developing programs at insurance companies, I found the actuarial teams were asked what they considered statistically sound data. This is what I heard: For auto, they needed 20,000 data years, which meant 20,000 cars. But on the home side, they were like "We don’t even know—maybe double that amount."
Don’t overlook customer-segmentation strategy
Dawn: When it comes time to find your test market, look at your book of business and think about ways to leverage data analytic models to identify your early adopter, tech-savvy, and do-it-yourself-type customers. Through understanding the purchasing decisions of customers, an insurer can target the messaging and invitations to what the customer is focused on. For instance, people who purchase items from big-box home improvement stores for home remodels may be very interested in water and smoke/fire sensors. Since they have just invested a significant amount of their time and energy in fixing up their home, they will likely care about keeping it safe from incidents.
Do focus on customer engagement
Drew: Carriers are realizing that if they get involved and change the relationship with their customers, then it’s not an adversarial relationship but truly a partnership. One thing carriers are doing is enabling their contract partners in their home repair network to perform proactive services. We see them partnering with companies like HomeAdvisor or TaskRabbit, which have applications that allow you to schedule home maintenance tasks. They’re making it easier for customers to get that gutter cleaned or that debris picked up, or to take care of anything that’s going to eventually cause more headaches for the homeowner.
Do think about your partnership strategy
Drew, I know you’ve managed insurance relationships while working at a smart home company. What’s your perspective on structuring a pilot from that vantage point?
Drew: The reality is, in the near-term, these technology companies are interested in unlocking new channel-distribution opportunities and hoping that these pilots prove to be successful and, most importantly, scalable. These early relationships are great, but they require a lot of time and energy!
Recognize that in many cases these are tech start-ups who are much smaller than you. Every time you make demands for what you would consider very standard privacy and security policies and practices, understand that can be a big ask. It doesn’t matter whether you’re talking 500 units or 50,000 units—the amount of work for the smart home company tends to be about the same. So as these brands assess where they’re going to plant their seeds, the obvious choice is to skew towards the insurer who wants to commit to larger volumes. It’s helpful when the insurer is upfront and realistic about what they want, and can say, “This is the volume commitment, this is the type of data we want, this is the cadence we’re going to want it at, and this is how we need to collaborate on things like the customer experience.”
Dawn: The challenge is that the smart home companies can only manage so many relationships at the same time. Imagine you’re a provider involved in pilots with several large insurance companies. They all want different data frequencies and different volumes of data. They want you to create a landing page for them. They want their customers to get different versions of email messages. They each have their own set of security requirements. How many pilots can one technology company manage?
Don’t forget there’s more than one IoT brand
Drew: We’ve talked about pilots as a starting point to test the waters, but let’s not lose sight of a likely future that has a range of different technologies, provided by a wide array of smart home brands within a carrier’s book of business. Carriers need to be thinking about how to unlock actionable insights at scale, independent of brand. How do you build operational processes to get that data into the hands of the people who are going to write that new business, renew that policy, or pay that claim?
Dawn: That’s where the Verisk Data Exchange solution is ideal. It solves this “many-to-many problem.” And it also addresses the challenge of operationalizing workflow data, because we’re going to integrate the data with the core systems insurance professionals use every day.
Drew: Regardless of where the data comes from, we can normalize it to provide meaningful, actionable insights. I advise looking at the data you get from smart devices as agnostically as possible. In your pilots, focus less on brand and more on sensor capabilities. If you’re solely dependent on a single partner or handful of partners, you’re leaving a lot of actionable data on the table.
Dawn: That’s certainly true, Drew. I think if you’re an insurer, the essential benefit of the data exchange is that you can get out of the game of worrying who to partner with, or what technology to deploy. You can move on to being purely data driven in how you operationalize or create innovative insurance products.