Business intelligence (BI) solutions must meet businessneeds to sustain sponsorship, funding, and support. However,meeting the business needs should not come at the expense ofproviding an infrastructure with the controls, alerts and workflowprocessing reports that increase the trust that the users have inthe solution and underlying data. These types of controls, alertsand reports often fall victim to meeting tight budgets andtimeframes of BI projects.

BI solutions and applications should be embarked upon to meet abusiness need. Over the years insurance companies' needs haveincluded reducing claim costs through management of claim adjusterworkloads, better pricing decisions through monitoring theapplication of base rates and scheduled modifications, andoptimization of producer channels through more completeinformation.

An empirical mark of a successful BI program is that itgenerates the demand from the business users for more data and newcapabilities. As the questions of today get answered, the questionsof tomorrow are uncovered. Many companies begin their journey bydeveloping roadmaps for their BI program that includes the types ofbenefits they are looking to achieve, when they will achieve thosebenefits, and how they will build out their BI capabilities in aniterative fashion while delivering upon those needs along the way.Often, these roadmaps will even show the points at which specificdepartments and executives will benefit.

To ensure business involvement and sustain funding, BI programslook for sponsorship and champions at the highest levels of theorganization. This is, especially true when the BI initiative isbeing driven by IT with the hope of positioning IT as a moreresponsive partner to their "customers" - the business users.

Historically, it is not uncommon for the business and ITmanagement to feel or perceive that there is a non-responsivenesson the part of the IT department. For instance, it is common tohear that a business manager or department has requested a reportand then waited weeks, if not months, until that report wasprioritized highly enough within IT to get it developed anddelivered.

The hope is that with (BI facilities) such as a data warehouseand data marts in place, reports can be built more quickly and thatthe business users can use the BI tools in a self-serviceenvironment, creating their own reports and even performing ad hocqueries on their own.

What is often left unsaid is that it is important to get highlevel executive sponsorship and champions to sustain interest andfunding for a project that will take considerable time and money todeliver - even for the earliest beneficiaries. In two decades ofdelivering reporting, data warehousing and business intelligencesolutions, the following scenario has played out time and timeagain.

The project starts with a lot of fanfare including managementand executives from multiple departments voicing their informationneeds. Next is the prioritization of those needs and identificationof the data that is required to meet those needs. At this point,there is a lot of excitement, from the business side of the houseas well as the technologists, about the promise of businessintelligence and the new capabilities that will be madeavailable.

Then, from a business perspective, progress seems to stop whileIT spending increases for anywhere from eight months to two yearsor more while the developers model, acquire, edit, organize,standardize, and transform the data into the single version of thetruth.

Many BI projects have either been cut after the initial phase orhalted during this period as the costs drag out and IT is seen,once again, as being non-responsive to the reporting needs of thecompany. One reason this happens is because the business problemschange between when the project began and the first delivery of asolution.

Another common reason for IT being seen as non-responsive ortaking too long is that interim, temporary solutions have alreadydelivered much of the value at lower cost and in less time. OtherBI projects have cut corners during delivery of the firstiteration; often resulting in dissatisfaction on the part of thebusiness users ending with the BI program being cut.

There are a number of reasons that IT groups try to cut cornerson early iterations of BI programs. Most often corners are cut inan effort to get the data in the hands of the users morequickly.

One area often cut is in the balancing and controlling of thedata. Balancing and controlling is the process of proving that datais not lost or altered in a financially material way from thesource to the reports. When balance and controls are cut, it isdifficult to prove that the data is correct and complete, leadingto mistrust of the data and a lack of adoption of the solution as awhole.

Another area that gets cut is "operational reporting" about theprocessing of the BI solution itself. Some of the BI operationalreporting has to do with balance and controls, but it also includeswhat data was brought in when. Whether or not the BI processing wassuccessful, it is important that the business feel that if theyhave a question about the numbers, they have somewhere to look tosee if the most recent information has been successfully broughtinto the data warehouse or data mart.

In a number of cases that I have seen, the calculated metrics inthe BI solution were proven to be more correct than the existingreports to which they were being reconciled; however, because therewas no overall reconciliation of the numbers in the BI solution andthe processes that created them, the users chose to continue withthe existing reports and not trust the new numbers coming out ofthe BI solution.

This then, is the current frontier of business intelligencetechnologies--to reduce the time it takes to create the singleversion of the truth while still including balance, controls andadequate processing information and, ultimately, to increase thetrust the users have in the solution and the data. ETL tools haveautomated some of these processes. However, with the proper tools,today it is possible to reduce the time it takes to make dataavailable in a BI solution that includes management and controlreports related to the processes involved in acquiring the data,managing data acquisition from new sources as they come on-line,and requiring minimal coding efforts. When it is possible to getdata in the hands of users in a matter of weeks instead of months,there will be fewer BI projects that are perceived as failures.

New BI technologies must make it easier to start small, growingincrementally with rapid turnaround on enhancements, additionaldata acquisition, and additional business applications. Tools thathelp IT to deliver a solution with balance, controls and processinginformation in a timely fashion increases the likelihood of successand adoption of the BI solution and in turn makes IT appear moreresponsiveness to the business.

Aviva Philips ([email protected]) is the seniorbusiness architect for insurance at Advanced System Designs. Shehas over twenty years experience delivering business intelligenceand business performance management solutions to the insuranceindustry.