«

»

Report Analytics

Share

Report analyticsReport Analytics scrapes data from existing reports into integrated reports for decision-makers. It is positioned as self-service business intelligence that eliminates the delays and costs of traditional data warehousing. However, I have a concern with the trustworthiness of the existing report data.

In a TDWI white paper Report Analytics was positioned as the new solution to turn disparate and diverse data into intelligence where conventional BI does not completely meet business needs.

Report Analytics supposedly allows end users to easily incorporate data from any combination of reports inside or outside the organisation, and then publish dynamic, interactive reports, without any IT involvement. So they bypass the data warehouse and all the complexities and delays it introduces. Report Analytics do not require changes to existing reports or processes, and no additional programming or databases. They claim that the accuracy of the data coming from existing reports need not be verified.

That exactly is my biggest concern. In many large data warehouse and business intelligence implementations, we have over-spent many hours to reconcile the new dashboards with existing reports. Most of the energy goes into explaining why the existing reports are inaccurate or incomplete. The reports often don’t cater for the latest products or client types, or apply business rules and criteria that are not valid any more. Now you want to blindly take those values for granted?

With a good ETL framework, developing the data warehouse and ETL processes aren’t time-consuming. Proving step by step to IT and the business that the source-to-staging, staging-to-warehouse and warehouse-to-dashboard transformations are “more accurate” than their existing reports are the real time killers.

A properly conformed data warehouse integrates data across business functions at the most granular transaction level. Scraping data off existing reports would typically integrate it at the department or daily summary level, which makes detailed drill-down impossible.

For me, Report Analytics can quickly demonstrate potential value to the business, and is very efficient for visual detailed requirements gathering, using realistic data values. For full-scale production reporting across diverse business units and their systems, I will still bet my bottom dollar on a properly conformed data warehouse. Running many different reports off the data warehouse will always give consistent results – that would be hard to guarantee with Report Analytics.

Leave a Reply