The insurance business has historically been an early adopter of new business intelligence (BI) solutions such as data warehousing, advanced reporting systems and data mining, as well as advanced statistical applications. The reasons for this advanced IT development in the insurance business are mainly due to a complex business model in a highly competitive environment. The insurance business has, at least until recently, been a sector that has had the money to invest heavily in new and often expensive IT solutions. The insurance business sells products where many factors must be taken into account to ensure profitability. Therefore, the analyses needed to manage the risks connected with insurance or funds tend to be highly complex.

Once the analyses are complete, a top priority for management in the insurance industry is often the need to implement and improve so-called dashboard or cockpit solutions. These applications are meant to help to understand the advanced analyses and also to communicate the results in an unambiguous manner.

What Is Peculiar About the Insurance Business?

The insurance business' needs for BI solutions are often different from what is included in the more-or-less standardized BI packages. Whereas the IT department in the average brick-and-mortar retail business may be approximately one percent of costs and staff, the figure for an insurance company is 10 percent. This is due to several reasons, each indicating special needs when it comes to analyses and presentations of data. The main differences are that the insurance business:

  • Does not have physical products.
  • Has products that often stay active over long periods to the same customers, indicating customer relationships that can span over 30 years.
  • Has products that are often highly personalized and that need to be maintained over long time periods.
  • Is very influenced by external factors.

Retail businesses have physical products that they deliver to their customers. They may have optimized their stock inventory, minimizing what they keep on the shelves; but, in the end, the customer gets a physical product.
On the other hand, insurance, just like banks and other financial services, offers the customer a piece of paper as the only directly visible product. This paper guarantees money according to certain predefined circumstances (such as an accident or, in happier cases, retirement). In other words, the value of the paper is that it guarantees possible future compensation. As a claim may be made 20 or 30 years after purchase, it is unlikely that the person who issued that policy will still be around remembering the relevant details should there be a problem with the stated data concerning the terms and conditions. This puts a tremendous demand on the exactitude of the stated information in the documents. Another consequence is that the volumes of data become enormous, as active products may be kept in the systems for as long as 40 years. On top of this, the data is most often spread throughout numerous applications.

In practical terms, this obviously means that the need for correct data management becomes extremely important, something that explains the considerable role of IT in financial companies. Also, dashboard solutions become of greater importance to the insurance business than most other businesses, as they present a rapid visual overview for voluminous and highly complex underlying data, often spanning long time periods.

In the insurance business, life insurance may not be profitable until the policyholder has made payments for 15 years. Added to this long-term management of products and customers is the number of future claims and their costs that must be estimated. For a property insurance business, some of these estimations are very volatile, as one hurricane can blow off all estimates and turn a forecasted annual profit into a loss. The long-term planning is made yet more difficult and important by the fact that many insurance companies offer products that guarantee a certain return to the client after a certain time, such as retirement funds. As we have all seen by now, it is impossible to predict the stock market (where some of these funds are invested); therefore, any future planning is even more uncertain.

The difficulties inherent in predicting the future put high demands on the analyses, whether they deal with financial risk, a future demographic situation or other factors. This, in turn, clearly explains the importance of statistical packages and why insurance companies are early adopters of data mining solutions.

Practically speaking, these complicated analyses over time periods spanning up to 30 years into the future not only need to be calculated, but also presented in a clear way. A typical analysis is the embedded value model.1 Even if the evolution of the embedded value is the result of some highly complex calculations, its purpose is simply to show if the value of the business is increasing or decreasing over time. Because of the complexity, it is necessary to use simple, yet powerful, presentations in order to show the estimated value of the business. Simple bar charts are often used for this purpose as they give the overview needed and are easily understood.

Insurance products may be highly customized, depending on their type. Some insurance products, especially for companies as opposed to private customers, can have as many different versions as they have customers. This means that the understanding of a specific product type can rapidly become very complex, as it may contain an almost unlimited number of attributes with different terms and conditions.

Highly customized products are certainly not unique to the insurance business; however, the lifetime of an insurance policy distinguishes it from most other products. Over time, the number of different products, all of which stay active, can become staggering. This complexity certainly demands optimized ways to present the actual business situation so as to not lose the overview.

Depending on the country, the insurance business is often highly regulated by the respective governments. These regulations may state, among other things, how money can and cannot be invested, what the interest rates are and what guarantees on return need to be given to the customers.

It is also often implied by law to provide certain statistics to governmental organizations. Although this fact does not necessarily directly affect the presentation of information, it will affect the data management and the analyses. Some data extractions and analyses are developed simply because they have to be delivered to outside governmental organizations.

BI Needs For the Insurance Business

As has been noted, the complexity of the insurance business puts high demands on how its activities are analyzed and presented. It can be said that the more complex the business gets, the easier the understanding of the presentations needs to be. Practically, this has manifested itself in the wide acceptance of dashboard or cockpit solutions in the insurance business. Simple charts, sometimes only showing red for bad or risky performance and green for satisfying performance, are often used to communicate the state of the business. These basic presentations are often enough to get the overview needed, regardless of the complexity of the underlying analyses.

The complexity of the insurance business also puts high demands on the unification of the business language (e.g., the definitions of different kinds of premiums). Here, it is of great importance that the actuaries agree with sales staff who agree with the accountants, etc.

The basic method for presenting complex analyses in a dashboard solution is straightforward in theory: define the analyses' key metrics and their meaning, set the goals and finally measure how actual results fare compared to the goals. In practice, the difficulty lies in defining what to measure, how to access the necessary data and how to ensure flexibility in the system so that it can adapt to new demands. These new demands may be both internal and external.

Given the complexity of the insurance business, the need for dashboard solutions for presenting and following the activities becomes quite obvious. However, because of the large amounts of data to analyze, data visualization –­ often directly linked to the dashboards –­ is often an excellent complement to both dashboard and data mining solutions. Data visualization applications used for dashboards add the ability to quickly obtain the underlying detail, should it be requested.

The dashboard presentations should be simple because of the underlying complexity. However, with many of today's software packages, these presentations can differ very much in a positive way. Anything from bar charts to visual thermometers can be used to show the results of the analyses or activities.

Considering the peculiarities of the insurance business and its complexity, paired with the obvious need to present this complexity in an unambiguous and comprehensive manner, the actions needed are often different from other businesses. The insurance business needs to be creative, something that has often manifested itself in its eagerness to adopt new IT solutions. At the same time, it could be argued that insurance does not have a reputation as innovative and cool to many people, something that is most often not true when it comes to its IT applications.

However, considering the competitiveness and the relatively stiff market compared to most non-insurance businesses, the need to be creative becomes increasingly more important. Given the need to analyze and present a highly complex and long-term –­ and, therefore, unpredictable –­ business, it is not enough to have the right tools. It is also necessary to allow a creative environment where the employees are able to try new ways of working. (A lack thereof may be why the insurance business has gotten a reputation for being anything but cool.) The challenges of the insurance business are in any case formidable, and the possibility to create innovative solutions is anything but uncool.

1. Embedded value is an integral part of today's insurance business. It is an estimate of the total value of a life insurance business and, in some cases, non-life as well. For this estimation, the embedded value model looks at the present insurance portfolio, excluding any value that may come from future businesses (sales, acquisitions, etc.). The calculation of the embedded value is complex, and there exist several different methods for computation of the value.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access