How Understory builds better weather insights

Weather data and insights is a huge business. IBM paid an estimated $2 billion for The Weather Company in 2015. In 2013, in the agriculture world, Monsanto purchased Climate Corporation for $930 million to get better weather data for crop insurance. Meanwhile, there are dozens of startups trying to use connected sensors and advanced modeling to build better weather insights.

Understory, which just raised $7.5 million, is one of those startups. The Madison, Wisconsin-based startup was formed in 2012 and built its business around its own proprietary sensors.

The Understory sensors are built using a ball on the top that can sense raindrops, hail, and wind as they hit the sphere. The sensors can tell how much rain there is, the speed of the wind, and the force and size of the hail based on those hits. Alex Kubicek, CEO of Understory, says this mechanism works much better than traditional rain gauges or anemometers that measure wind. That’s because, over time, rain sensors can get clogged with debris, while anemometers can get broken easily, especially after ice storms or if they’re hit by chunks of hail.

Kubicek says Understory’s sensors are relatively cheap to build, are solar-powered, and come with their own cellular connection. In cities, Understory does deals with building owners to place the devices on roofs so it can establish networks in specific areas. For example, in Dallas-Fort Worth, Understory has a sensor placed every two miles so it can get granular data for insurance companies.

Insurers there want to understand how hailstorms might impact local policyholders. So an insurer might use data from Understory to reject claims related to a hailstorm that didn’t register on the Understory sensors. Or it might use the Understory data to see if the insurer should prepare for abnormal losses. In addition to insurers, Understory provides the data to the region for its own research and emergency preparedness efforts for free.

Not every sensor network has to be as dense as the Dallas-Fort Worth network, says Kubicek. It depends on what the data will be used for and what type of weather the area has. For example, the Texas sensors need to be close to one another because the intensity of hailstorms can vary over a wide area.

However, in Argentina, where Understory is building a network so it can provide weather and plant data to Monsanto, sensors are placed further apart from one another. There, the sensors are looking for humidity and temperature data to augment the local weather service, which isn’t very accurate.

As someone who watches the internet of things, I think Understory’s biggest differentiator is its sensor. Broadly speaking, there’s a real question as to how a company can differentiate itself when data seems to be everywhere. Building a dedicated specialty sensor is one way, but in many cases such an effort makes the sensor so expensive it doesn’t get deployed. And the market winner usually turns out to be the company that compensates for cheap, generic sensors with better algorithms and more data.

However, if a company can design a sensor and get high-quality data that comes from a wide enough network of those sensors to be useful, it does have a real advantage. Especially in weather, which can vary significantly from one mile to another. Weather is a huge factor in everything from agriculture to shopping habits, so there are plenty of customers willing to pay for high-quality weather data.

So far, Understory is focused on property and casualty insurers and the agriculture market, but it could easily find more customers in the financial world or with utilities. Its big challenge will be building out a leasing infrastructure that lets it put sensors in more places — and making sure the models it uses to turn its data into insights provide real value.