Sutterworld Logi


Multi-employer and public funds have a long horizon for their data needs: systems change, plan rules change and years of contribution history for pensions and annuities must be maintained. Building a data warehouse is one solution to taking the data that’s trapped in your operational systems and making it useful. To free the valuable data and make it accessible will involve lots of time, people, resources, and possibly new technologies. But before getting lost in the language and technical aspects of data warehousing, there are five guiding principles I recommend for designing the data warehouse that can help your project be a success.

- Define your objectives. Designing the data warehouse can become a contentious struggle between business users and IT developers. Take the time to reach consensus and publish the goals of the data warehouse. This important step can serve as the guardrails that keep the project on track.

-  Design for the users. Be prepared to answer design considerations for how the users will use the data. You want to be sure that the design supports users in asking the right questions to get accurate answers.

 -  Design for the benefit plan. The designers of your data warehouse schema will need to understand many of the fund’s legacy business rules and other business processes that generate data. Every fund office is different and understanding the rules is just as important as understanding the technology.

-  Design in stages. Often, organizations that try to design a warehouse for the entire enterprise will encounter rework along the way. Business processes or project dimensions may change over time before the whole warehouse is deployed. Designing in stages by subject areas (claims processing or pension contribution history, for example) will be more efficient and expedient than trying to design for the entire enterprise.

-  Design for Usage. Form follows function when in comes to historical data sets. Keep in mind that in some cases, you’ll need to deliver information extremely quickly. Form usually follows function so designing de-normalized and summary tables may be the best approach to support rapid response time. Indexing methods or specific hardware appliances can boost performance of queries. Almost all considerations map back to knowing how the data will be used.

Once you have designed and implemented the data warehouse, the work doesn’t stop there. Be sure to review the original goals defined for the data warehouse and keep a record of fund office business decisions that are made as result of data warehouse queries. That way, you can justify and understand the value of the warehouse and learn from the decisions made along the way.

John Sutter
John is a Partner at MIDIOR, where he is a practice lead for both management consulting and technology services. With a background spanning systems architecture, user interface design and data analytics, John is known for helping clients change the way they think about data and most recently has focused his attention on mobile applications and analytics to support claims auto-adjudication.ew Paragraph
Share by: