Archive for June, 2011
Getting IT out of the report writing business
The creation and maintenance of Analytical Reporting has long been the bane of IT and with good reason. Reporting is necessarily reactive to the business, end users are often unsure of what they really need (or currently have for that matter) and increasing data complexity makes it time and labor intensive. This makes reporting nearly impossible to forecast from a budget or resource perspective.
In the last few years, both existing and new reporting/analytic tools have become much more user friendly. True user self-service is now a reality across the spectrum of price points. At the high end of the spectrum MicroStrategy, Business Objects and Cognos are still getting stronger, especially on their analytics capabilities. What is truly remarkable are the features found in what might be labeled midmarket products. Tableau, Qlikview, Pentaho Enterprise and Jaspersoft Enterprise all over very strong user self service at entry-level price points. While each has its strengths, all are potentially viable tools for the large enterprise.
Leading the horse to water
In our experience, most business organizations respond positively to the self-service concept. Sticking points with the business usually involve either the budget or availability of the expertise necessary to take on self-service. The sticking point with IT is building a consistent and reliable infrastructure foundation for self-service.
Data governance is key to making self-service a reality. By data governance we simple mean data accuracy and integrity. While each business has a unique set of issues, users need to be notified if upstream processes have failed. Another crucial aspect of governance is data quality. It is difficult and prohibitively expensive to fix upstream data quality processes in the analytics layer. Data quality needs to be addressed where the quality problem is created. As data is added the business and IT need to revisit the environment to ensure quality is maintained.
User self-service can go too far
The goal of self-service is to balance cost, efficiency and responsiveness of analytical reporting to a business. It is possible to go too far, These new analytic tools also have ETL functionality and the ability to proliferate data cubes, thus creating data warehouses on top of data warehouses. This approach ultimately “complexifies” the analytic environment creating expensive and unresponsive overhead. An on-going data governance process will allow the organization to anticipate and manage increasing data complexity.
Part 3 of 3: Decreasing your total cost of ownership
At the end of March, Rita Sallam from Gartner published the report “BI Platforms User Survey, 2011: Customers Rate Their BI Platform Vendor Cost of Ownership.” A fifteen minute survey was completed by 1,225 BI professionals across the globe, of which 20% didn’t come from a vendor reference list. We thought it would be helpful to interpret the results and include our insights from being on the ground for several implementations.
If you would like to DECREASE your total cost of ownership, the following tidbits of advice will help justify your Business Intelligence investments:
5. Manage process for business ad-hoc reporting requests
Questions about current BI environment: Do your business users deploy their own technology rather than going through the proper IT channels? As a result, do business users create their own metadata, degrading data quality?
Suggestion to manage ad-hoc reporting requests: Create a self-service architecture that allows IT to get out of the report writing business.
6. Accurately measure total cost of ownership at beginning of implementation, and manage expectations.
Suggestion to accurately measure TCO: Look at both hard and soft costs of the initial implementation, support and maintenance. Include number of users, skill level (super user, part time user, etc.), complexity of data, size of data, cost of rolling out, time and effort for new additions, IT reputation and business sanity.
