For utility and energy companies considering a business intelligence solution such as a data warehouse, here are ten tips that will help to ensure a successful implementation: 1. Involve the Business Users.
Get the users involved early on in the project and regularly throughout development. For larger companies, consider forming a core team of strategic users that have the vision to know how the data warehouse will benefit the company and can make rational decisions related to scope and priorities. Even if your company currently has a data warehouse or data mart, interview representative users from each functional area to identify significant changes in the decision-making process. Use the findings from these interviews to challenge each of the original requirements for the data warehouse.

2. But, Don't Expect Them to Have All the Answers.
Be prepared to face users that do not fully know their requirements. There are a number of reasons for this: the business may have recently changed, new processes may have been put into place or new roles might have been assigned. On the other hand, users may simply not understand the concept of a data warehouse or other analytical store of business data. Thus they can't verbalize how they would like to see the information. Even if users appear to know their needs, be prepared for change. Better yet, embrace it! With deregulation, new products will be introduced, others will be terminated, competition will be an everyday reality along with pressure to maintain cost control. Users' needs can fluctuate wildly throughout this process.

3. Evaluate All Possible Sources of Data.
Take an inventory of all possible data sources throughout the enterprise. Include in the inventory the type of data that is being captured, data quality, platform of the system, format of the actual data, timing of when it is updated, optimal time periods to extract data and any other issues related to accessing the source system. Some of the "systems" that may be included in the inventory will actually be spreadsheets, personal desktop databases or hard copy reports. Be sure not to overlook pending enhancements to existing systems or near-term efforts to completely replace a current system.

4. Lay Out a Long-Term Architecture.
Take the time to design an overall architecture for long-term decision making in the enterprise. Leveraging the findings from the user requirements gathering step and the inventory of data sources, the architecture should be a master plan to guide phased development of the data warehouse. For example, for a utility that will be creating a data warehouse consisting of subject-oriented data marts to address customer billing, meter readings, meter inventory and marketing campaign analysis, the architecture might include the data sources needed to populate these data marts, a staging area to store the component extracts and conceptual star schemas for the data marts that identify dimension tables shared across the warehouse. The architecture will be a valuable resource to refer to during development, and it will ensure that the data marts are integrated in such a way to allow for analysis across their data sets. A good architecture is one that provides a framework for flexibility and scalability.

5. Iterate Through the Design.
Incorporate an iterative philosophy that offers incremental functionality to business users. A reasonable goal is to deliver the first production-ready iteration in three months. This will give the development team an opportunity to confirm with the users what they understood to be the requirements. For the users, particularly those new to decision support systems, they will have something tangible against which they can more precisely define their requirements. Avoid the "big bang" approach that takes many years to complete and leaves your users and your CFO empty-handed.

6. Seek Out Third-Party Data Providers. The development team should seek out third-party data providers that sell information on competitors and the overall industry segment. For energy and utility companies, third-party data may include:

  • Electrical or gas usage for target customers,
  • Spot prices and futures in other regions, and
  • Weather information.

What gas marketer wouldn't want to see current customer gas usage, along with weather data and spot prices, all in one report?
7. Embrace the Customer.
At least one subject area within the data warehouse architecture should be focused on the customer. Ideally, a majority of data marts will share the same customer dimension, and thereby improve the consistency of viewing and reporting about customers. With deregulation, the need to satisfy the needs of the customer is even more paramount. Data warehouse users must easily see all the usage, billing and marketing activity by customer. When customer hierarchies exist, such as parent company, division and subsidiary, all of these levels must be accurately modeled in the data warehouse to provide a consistent view of the customer.

8. Intuitive User Access.
With new products and new sales channels being offered, a broader user community will emerge that is less technical and more business focused. Existing access tools in the data warehouse may need to be augmented or replaced with intuitive, graphical front-end applications. These access tools must be able to support anyone from the casual user who simply reads and refreshes reports to the power user who frequently creates unstructured, ad hoc queries of the data warehouse. Don't be disappointed if you can't find one tool that fits all. You might consider offering two standard access tools to users ­ one for casual users and another for users needing powerful analytics. However, if you offer more than two front ends, you may find your support staff growing to match the size of the total user community!

9. Customer Access.
Once you have your data warehouse up and running, be prepared to hear from customers who want to access it, particularly the large industrial or commercial accounts whose energy bills are a significant chunk of their operating costs. Other customers that may want access include marketers of gas and electricity, regulatory agencies and savvy residential accounts.

10. Keep it Simple.
With all the changes in the industry, it is imperative that the data warehouse team deliver an application that is simple and intuitive ­ both for the users to access and for IT personnel to maintain. Scrutinize the many packaged data mart vendors to find those with an impressive list of paid, installed sites. These tools typically consist of an integrated suite that handles the three key architectural components of a data mart: extract/transformation/load (ETL), database repository and front-end query/reporting. As these vendors continue to refine their suites, customers will stand to benefit from the increased simplicity.

If you are the CIO, IT director or the executive business sponsor and you are launching a new data warehouse effort (or perhaps trying to rekindle a stalled effort), encourage your development team to follow these ten guidelines. Also, visit with other similarly sized companies that have had a data warehouse in production for at least one year. They will fill your ear with more valuable insights. Your customers will be pleased and your bottom-line will improve.

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