Big data adds new flavors to plain-vanilla auto-insurance By Kishore Jethanandani Big Data is helping auto-insurance companies find unfamiliar and varied sources of revenue from services to mitigate risk in driving. Currently, most insurance companies look to corral relatively safe drivers who account for a fraction of the potential of the larger market for auto-insurance. Big Data opens a wider market with measurement of patterns of risky driving behavior and its modification for the large majority of drivers. Cloud-connected telematics devices enable insurance companies to gather data on vehicle usage. They can analyze the data in real-time and send prompts to drivers when their behavior is susceptible to property damage or bodily harm. Data gathered by Allstate with the help of its Drive Wise device, in the states of Illinois, Ohio and Arizona, found that the percentage of drivers with acceptable levels of risk rose from an initial 25 percent of the sample group to 75 percent at the end of the trial according to data reported by Insurance Networking. This was achieved by identifying driving habits, based on data of 11 million miles of driving over 350,000 hours that predicted low loss outcomes. The drivers who conformed to safe driving behavior are rewarded with 10 percent discounts which rise to 30 percent over several terms. Currently, insurance companies use thumb rules to determine their premium levels for sub-prime customers. Teen drivers, for example, typically pay higher premiums as they are all assumed to be unsafe drivers. All too often, teens report themselves as occasional drivers on their parents’ insurance plans to save on the high charges and insurance companies lose their business.
A segmented approach to insurance plans is illustrated by American Family Insurance which has a Safe Teen Driver Program designed to train young drivers to learn safe driving. It has a device that automatically captures video of undesirable events such as a crash or a sudden turn around a corner. The video is captured a few seconds before and after the incident. This footage is reviewed by a driving coach who advises a plan to change driving habits for the teenager involved. This program has reduced risky driving behavior among teenagers by 70 percent according to a report by Transport Business. The continuous inflow of data enables insurance companies to reassess risk behavior at frequent intervals and adjust premiums as drivers improve performance. Location is another variable that affects the risk of insuring vehicles and is best assessed with data on the routes travelled by them. ISO, a member of the Verisk Analytics Group, gathers the data from insurance companies in the industry and estimates the average risk of driving in each of the regions in states across the USA. Clients can expect discounts of up to 25 percent depending on where they drive. SaaS-based fleet management companies like Telogis are best equipped to gather the large volumes of driving data especially for commercial fleets. Its platform records routes travelled as well as risky behavior like hard braking, sudden turns and unusually rapid acceleration. The sum total of the data is aggregated in a scorecard that is communicated to supervisors in real-time. Auto-insurance markets stagnated with the mispricing of risk that was rampant when the data available to insurance companies was scarce or hard to analyze in sufficient detail. SaaS-enabled Big Data analytics lets insurance companies price
for value and makes it attractive for more customers to buy insurance plans. Risk mitigation lowers costs and expands the size of insurance markets. Big Data Spawns Farm Insurance By Kishore Jethanandani For the legacy Federal Farm Insurance, the backdrop has been the vagaries of farm life described in the seminal novel ‘Grapes of Wrath’. Federal Insurance for agriculture is a one-size-fits-all plan while the Total Weather Insurance Plan of the Climate Corporation, a Big Data SaaS company, customizes an insurance plan for each farm. Insurance coverage, under the Federal Plan, kicks in when the output falls below 120 bushels per acre. It tends to willy-nilly discriminate against the more productive farmers whose baseline rate of production is higher and may not fall below 120 bushels except in the worst of circumstances. Beginning from 2012, the Federal Insurance Plan has been modified, for some counties and crops to factor in the average growth in productivity. It still discriminates against the most productive farmers who may be risking higher investments in nutrients and technology to increase productivity. By contrast, the Climate Corporation sees weather as the primary determinant of variability of farm output. Insurance is paid out automatically, without a visit from the inspectors, in the event of any deviations from expected parameters of weather. The Climate Corporation estimates the risk for individual farms. It crunches the 60 plus years of data for crop yield from the Department of Agriculture, 14 terabytes of it, and matches that with weather data from one million points the government gathers with Doppler radars. The expected yield data, based on weather and soils conditions, for every two square miles of farm land in the USA, is used to estimate the potential exposure as a result of any variation in the weather. This is used to price the premium for each farm.
Estimates of the risk not covered by Federal Insurance vary in the range of fifteen and twenty five percent. This is often amplified by price variations that have been more common with the gyrations in the global commodity markets. The numbers are far higher for the most productive farmers in the ballpark of 60-70 bushels per acre. The data storehouse with the Climate Corporation has more than one uses. It is also used for risk mitigation and productivity gains. Farmers can use interactive maps on their IPADs or computers to estimate their risks of planting based on forecasts and actual data on rain, wind or heat. In the Northern areas of the mid-west, for example, farmers struggle to make decisions for a second crop, typically soybean, because an early freeze damages crops. Back-of-the-envelope estimates of weather conditions are not good enough or county-wide forecasts are not useful for individual farmers. The Climate Corporation prepares detailed estimates of the risk of early freeze for a thousand locations. Similarly, farmers in the Corn Belt struggle to make decisions about early planting to avoid the heat when pollination happens. In the North-Eastern belt farmers risk encountering freezing weather and risk to germination while the western states of the mid-west region are more likely to benefit from early planting to avoid the heat later in the season. Big Data is not just for the Big Corporations. It can be accessible to even small farmers who can plan their work to avoid unexpected losses from changes in their environment.