Insurance is a business with as many stats as baseball. We don'tmake products, we make promises. Those promises are recorded andstored as electronic data. In a very real and important senseinsurance is a data-based "industry."

And yet, until recent years, many carriers ran with minimalinformation about the effectiveness of their various businessoperations—underwriting, customer service, distribution or claimhandling. Indeed, companies struggled to produce basic financialreports, accurately estimate their claim liabilities or state theirreinsurance recoverables. The reasons for this sad state of affairswere several and included operational core systems with inadequatedata stores, multiple operational core systems with differing datameanings that were difficult to reconcile and balance, multiplemanual steps where data would be transferred and manipulated (oftenwith errors), and an almost complete lack of reporting andanalytical data storage and reporting.

Two major trends in insurance IT are transforming this picture.First, as we have discussed at length in this column, there is asustained movement across the industry to modernize core systems,often replacing aged systems that have brittle and restricted datastores with modern, flexible systems that store more informationand are better able to integrate data across platforms andsystems.

Second, business intelligence (BI) initiatives are appearingacross the industry that hold the potential to transform decisionmaking and operations management into an analytical process basedon historical, current, and predictive business insights. Not onlyare carriers getting better at capturing and storing moreinternally generated data they are also capturing and using datafrom external sources to supplement business decisions. Indeed wehave declared the arrival of the era of "big data", which certainlydoesn't mean it's here, but it is definitely on the way.

Carriers are aiming BI initiatives at various areas of theinsurance enterprise. Not surprisingly the two key areas that getattention are underwriting and claims, the two most significantprofit levers in an insurance company.

In the underwriting area BI is being used to analyze and scorerisks more accurately and to support the development of pricingmethods. These focus areas tend to interact and to feed off eachother. As carriers get more sophisticated at risk segmentation theyalso can get more sophisticated at pricing. As they know more abouttheir own profit profile across business segments they can morefinely adjust their position in the marketplace, in response tomarket pressures, regulatory changes, and company values.

BI may be more widely used in claims than inany other part of the business operation. Given that claims spends,roughly, seventy cents of every carrier dollar it is not surprisingthat carriers have focused on this important area, particularly oncauses of "leakage" including fraud, abuse, and inefficiency.

There are significant areas of claims that offer major savingswithout touching the sensitive area of indemnification, which is,after all, the "delivery of the insurance promise." Fraud, in allits diverse manifestations, is estimated to cost the industrybetween $20 billion and $30 billion annually. For certain carriers,e.g. workers' comp and (nonstandard) auto, fraud is a major issueto be constantly monitored and identified.

The identification of various patterns in the claims datauniverse can indicate fraud—multiple claimants on different losseswith the same home address; the high frequency appearance of thesame medical provider across numerous claims—can indicate areas ofSIU focus.

Another area of claims leakage is subrogation. Often carriershave only poor information concerning the circumstances of a loss,the degree of negligence across involved parties and the extent towhich recovery is available, or is pursued against another carrier.Estimates of subrogation recovery range across carriers, but ingeneral are in the high teens and low twenties as a percentage.Given that subrogation recoveries are "found money" it is an arearipe for the application of more data and some analysis.

A similar case can be made for salvage and reinsurance. Trackingthe application of reinsurance across losses and aggregatingrecoveries across multiple reinsurance treaties is usually a"spreadsheet function" that happens somewhat mysteriously in theaccounting department. Some carriers have extended the BIcapabilities to identify and take action in certain indemnityareas.

For example, one carrier undertook an analysis of the time ittook to schedule damaged vehicles for repair and concluded thatmajor savings could be made in rental reimbursement coverage—withan attendant increase in customer satisfaction—by getting vehiclesto approved body shops and therefore back to their owners morequickly.

Other claims-related uses of BI include claimintake and adjustor assignment, litigation management, andsettlement tracking and analysis.

All this activity has both attracted and been stimulated byvarious vendor offerings. As with other vendor segments there aresolutions that come from outside the industry and those that havegrown up within it. Several well known and long established namesdate back to the days of 4GL horizontal solutions and now playstrongly in the insurance space, emphasizing their deep roots in BIand their technology capabilities.

Then there are vertical market vendors that tout their industryknowledge first and their BI technology second. The appropriatenessof one vendor over anther, as always, depends on the needs andwants of any specific carrier. Some carriers are looking for toolsto build a BI capability, while others are looking for advice andguidance as to what should be in a warehouse and what kinds ofreporting can be done.

The vendor market segments into those solutions that provideboth insurance repository data models and reporting tools, or thosethat provide a framework for creating both data structures andreporting logic. Industry analyst Novarica segments the market intothree offerings:

• Platforms that provide tools for reporting and dataanalytics, and data storage and mapping. Platforms don'tnecessarily provide insurance content such as data models, standardreports or dashboards. Rather they enable carriers to develop theirown bespoke BI environment. Platforms are for carriers that for onereason or another want or feel the need to "build rather than buy"their BI solution.

• Models provide pre-build insurance data content, ETLcapabilities that extract insurance data from the operationalsystems environment and consolidate it into their pre-definedwarehouse schemas, which are in turn reported on and analyzed bypre-build applications. This option is more appropriate for thecarrier that is looking for significant guidance, and a tailoringproject rather than a build project in order to create their finalBI capability.

• Solution suites are a combinationof the platform and model offerings. The solution suite provides acomprehensive pre-built BI capability where the model and platformmay be tightly coupled and pre-integrated by one vendor, or maybemore loosely coupled.

Regardless of how much the carrier buys rather than builds theend result ends up structurally similar. The key components of a BIenvironment are:

• The ETL layer, which forms the bridge between thevarious underlying operational systems with all their differencesand vagaries and a single centralized data store that represents a"single version of the truth." As the mnemonic reminds us ETLprograms extract data from operational and legacy systems,transform it according to mapping rules into consumable formats andload it into a consolidated data structure, which may or may nothave been largely defined by a purchased insurance-specificmodel.

• The consolidated data store comes in at least twoflavors. One flavor is a two stage implementation where the firstdata store is an operational data store (ODS), which is intended toreplace operational systems as the system of record and mask theremainder of the IT environment from changes in underlying systems;and the second data store is the data warehouse, which is a versionof the ODS optimized for reporting and analytical work. The secondflavor eliminates the ODS and has individual operational systemsfeeding it directly.

The Reporting Layer usually consists of several capabilitiesincluding:

• Canned reports developed to meet regular datareporting requirements.

• Ad hoc reporting facilities where "power users"manipulate data cubes or Excel pivot tables to generate their owndetailed business insights.

• Higher level dashboards and gauges that providevisual and graphical representations of key business metrics andhigh level measures.

Taken together these capabilities are creatingintelligent companies and fostering a culture of enquiry andinsight. A common theme among companies that have made asignificant commitment to the BI world is that it creates its owndemand. Business users become more curious as they become moreinsightful and educated about the aspects of the business they areinvolved in.

New insights create new questions and new answers create greateraccuracy and efficiency. And like the competitive imperativecreated by the move to new core operational systems, the rise of BIhas created competitive capabilities that carriers have to match inorder to continue to compete. The days of running with the lightsoff and the dashboard blank are over; which is only appropriate foran industry that is in the data business.

George Grieve is CEO for CastleBay Consulting. Previously aCIO and still an acting consultant, he has spent much of the past25 years with property/casualty insurers, assisting them in thesearch, selection, negotiation, and implementation ofmission-critical, core insurance processing systems. He can bereached at 210-887-6423.

The content of "Shop Talk" is the responsibility of the author.Views and opinions are those of the author and do not necessarilyrepresent those of Tech Decisions.