When data and analytics change the game

Let’s face it. In many cases, analytics and data aren’t transformative for an organization. Sure, they enable better decisions, better offers to customers, better management of resources, and other “betters.” But they don’t change the game. “Moneyball” in baseball, for example, gave the teams who employed the analytical tools a small edge. If the tools were transformative, however, the Oakland A’s would have won a World Series within the last 25 years.

But sometimes data and analytics can change the game altogether. They can create an entirely new business model or approach to operations. They can allow organizations to engage with customers in an entirely new way, or bring radically new products and services to the marketplace. Of course, these breakthrough innovations don’t always succeed in changing the game, but it’s often worth the try.

I’ve been struck recently by several transformative developments in the underwriting area, for both business loans and insurance. One is actual, the other only partially implemented. But they’re illustrations of how companies can change the game completely with data and analytics.

I learned about an actual game-changer recently when I gave a presentation at Georgia Tech on big data and analytics, and attended a panel about their use in financial services. One of the panelists was the CEO of Kabbage, an Atlanta-based firm that provides business financing. Instead of asking you to fill out a long application for credit, Kabbage accesses (with your permission) your data from transactions on eBay, Square, Intuit Quickbooks, etc. Instead of waiting a few days or weeks to decide whether you are worthy of credit, it gives you a verdict in minutes. And instead of giving you a set credit line to be accessed for the foreseeable future, Kabbage monitors your data constantly and may raise or lower your credit line each day.

This strikes me as a huge advance over traditional credit underwriting. The decisions are more accurate because they use real data on business transactions. The customer service is better because the decisions are faster and less burdensome to the customer, and the lender is less exposed to changes in the borrower’s business because the credit decision-making is continuous rather than one-time.

The same data-centric underwriting approaches are also transforming property and casualty (P&C) insurance underwriting, particularly in auto insurance. Progressive Insurance has been marketing a “usage-based insurance” (UBI) approach to underwriting for about 15 years, and now there are competing offerings from such large insurers like Allstate, State Farm, and Aviva. UBI approaches, also known as “pay as you drive,” base their underwriting in part on how you actually drive (from data extracted from telematics devices), rather than segmentation or other modeling approaches that predict how you drive. As a Deloitte report on UBI suggests, the idea is rapidly catching on in the industry. Progressive now says that 30 percent of its new customers accept UBI (in exchange for discounts), and the company has accumulated more than 10 billion miles of driver data to analyze. The data lead to predictions of driving behavior that are substantially more accurate than any other predictor.

So, UBI is already a game-changer in this industry, but it has only begun to transform the industry altogether. One possible direction for the technology and related analytics is to offer value-added services to drivers. Once you have an accelerometer and GPS device in the car, for example, you can offer services involving maintenance (“the vibration pattern in your car suggests it needs front end alignment”), roadside assistance (“we noticed your car just broke down, and we have a truck within 5 minutes of it”), and geo-fencing services (“alert—your teenager just crossed the state line.”)

Another even more interesting direction involves changing the entire nature of insurance services. The goal of P&C insurance has always been to reimburse policy-holders for losses in case of an accident. But with data on how people drive, insurers could focus on accident prevention rather than reimbursement. They might say to policyholders (presumably to those who “opt in” to such arrangements), “We’ve noticed that you consistently drive 10 MPH or more over speed limits after 8PM. Our analysis suggests that this makes you 32 percent more likely to have a major accident than the average driver.” That might not change the behavior of risky drivers, but then again, such risky drivers would probably have been dropped by UBI providers before such a message was necessary. If a driver intends to decrease risk, this type of information could be very helpful in doing so.

As John Lucker and I point out in a Deloitte Review article, another type of data-generating device, the personal activity tracker, has the potential to transform the life and health insurance industries in a similar fashion. Instead of pricing health insurance based on predictions involving age, gender, smoking status, and so forth, insurance companies could incorporate actual physical activity into the mix of predictive factors. It’s been found in one study, for example, that walking speed is a good predictor of additional lifespan. To gather and report average walking speed is child’s play for an activity tracker. It would take a substantial change in the relationship between insurers and their customers to implement such an offering, but such programs are in place outside the United States. And very recently, John Hancock announced a new life insurance policy that monitors Fitbit activity data in exchange for discounted rates.

Of course, these industry-transforming applications of data and analytics raise interesting questions about personal privacy and the very nature of the industries involved. Perhaps substantial numbers of customers will rebel against such monitoring and eschew products based on it. Perhaps regulators will prohibit insurance and banking products of this type. To my mind, however, basing insurance and credit decisions on how you actually behave—rather than on predictions of how you will behave—is fundamentally fair. Furthermore, in all these cases I’ve discussed, the people and organizations that adopt them are likely to be the best customers and the lowest risks. Companies who pursue such customers, then, are likely to be successful.

This posting originally appeared in the Essays section of Deloitte University Press.
Copyright ? 2015 Deloitte Development LLC. All rights reserved.