Competitive pricing, stringent regulatory reporting and retention remain continuous challenges for insurers and, for these reasons and more, predictive analytics continues to be squarely in their radar screen as a critical differentiating capability. However, to date, few insurers have been able to unlock the full potential of this capability.

It is no longer a question of finding the right price for the risk. Risk assessment and pricing has evolved into a scientific subject area which at the same time continues to perplex insurers in this world of increasing uncertainty. Insurers need to look deeper and broader in their use of analytics for risk assessment as well as identifying, growing and retaining potential customers.

Insurers need to move from broad segmentation of risk such as age, weight or health risks (individuals) and industry, size or location (corporations), to a more granular level “test cells,” using many such factors simultaneously. While predictive scoring achieves this, it traditionally supplants and fails to leverage the actuarial experience gathered over decades. Insurers need to use a hybrid approach that combines scores and traditional actuarial practices. This is where predictive analytics comes into play.

Use of analytics needs to expand beyond risk management alone. One of the most underleveraged areas we found is in product development, where analytics is looked at in the context of underwriting. Other areas that are typically underleveraged are loyalty, cross sell/up-sell and enterprise performance management.

Property-casualty or life-health insurance may have different industry dynamics, but the need for better intelligence continues to be the same.

For life insurers, it may be 20%-30% cheaper to grow the portfolio through an acquisition than through a traditional agency channel. This has always been driven by the fact that life & health insurance is “sold,” not “bought.” Given the high expense of acquiring a new customer, two factors become critical for an insurer: better understanding of the key life events of a customer that warrant a sale, and proactively managing the retention of these customers.

Retention is no longer a given in life & health insurance. A good example is a product like term life, where a customer has the choice to shop for the best price through the push of a button. This has made the market fiercely competitive, causing carriers to think of innovative ways of designing the right product for the customer. An example can be taken from the credit card market where the players who have emerged as the leaders have perfected the use of predictive analytics to balance the customer’s price sensitivity with overall product profitability.

Furthermore, being able to manage retention requires establishing a process that allows an insurer to act before it is too late. In order to do so, we have found that it is important to look at three key areas.

First, it is important to predict which customers’ policies are about to end well before the event occurs. This predictive capability is typically very difficult to achieve with internal information alone. But by matching external information (for example, bureau data) with internal customer data, companies can identify those likely to drop out in enough time to permit the development and deployment of aggressive interventionist campaigns targeted to the sub segment of expiring accounts that are profitable and need to be retained.

Second, companies need to identify exactly what it will take to stop the customer from ending the policy. Once again, through the use of external information, companies can predict the actions necessary to change the customers’ behavior. Some of the reasons for a negative experience may include misunderstanding of what the insurer covered, versus what was paid. Turning this experience into a positive one requires creating alternatives for what the insurer can offer to this client in terms of coverage, price options, etc.

The third area is figuring out how to implement these findings.

With long term products such as life insurance and annuities, companies may be missing out on opportunities to capitalize fully on the changing life stages of their customers. For example, a 30-year-old person who purchases a 30-year term policy with a face value of $50,000, and “is just starting out his/her career” with a lot of debt and a modest salary, will likely have a completely different profile five years later.

By selectively pulling and aggregating external with internal data, companies can build highly accurate models that allow customized cross- and up-sell solutions to be developed for specific customer segments that exhibit the types of profiles of most interest (and profit) to the company.

For example, by tracking a customer’s, credit bureau history, an insurer can see how this individual shifts from a borrower (e.g., high credit card debt) to an asset builder (e.g., establishment of a mortgage and a home equity loan) to a saver (e.g., high credit lines, does not carry a balance on his/her credit cards, etc).

A new frontier is expanding into non-traditional products such as banking. Insurers are fully aware that there is a strong correlation between a customer’s credit risk and his/her insurance risk. Recently insurers have proved the same concept in reverse. Insurers have found that selling credit cards and other lending products to their most loyal and reliable customers has proven to be very profitable.

In summary, useful application of predictive analytics will be the driver that separates the leaders from followers in the industry. The application will span beyond risk, from product design and target marketing to customer servicing. Analysis will expand well beyond the use of internal data, into emerging and well-established sources as insurers search for the holy grail of predictive attributes that provides them with a true edge over their ever increasing competition.