People expect intelligence out of BI reports right off the bat. While it is a justified expectation, in reality, it requires a lot of effort, thoughts before getting actionable intelligence from BI reports. For e.g., Revenue Performance Report helps us to analyse Revenue flow for the past 12 months and also gives us the ability to drill down to month/ day level data to understand the various impact points. Representing the revenue performance data as a chart visually enhances the understanding of the report. If there is a revenue dip in the month of April, it is easier to identify and drill down to analyse the root causes. Reports like these need a lot of effort from the Business Analyst (BA) to identify the level of intelligence that should be built into the report and to ensure that the selection of right data to be extracted and displayed. The BI developers from their standpoint should do a feasibility study to assess the report requirements and then develop the reports.
In order to generate an actionable intelligence on the example report stated above, Trend on Revenue performance, three different teams need to work together:
- End users (marketing / sales / etc.)
- Business Analysts (BA)
- BI Development team
The BA can provide all the necessary inputs towards creating a data model, cubes, writing queries and for extracting data for the reports. Once the report is developed the BA can analyse and verify if it reflects the reality and then can apply the “what-if” rule to understand the impact of changing the report parameters. Until the reports reflect the reality, the BI developers will have to work closely with the BA for generating iterations of the reports, use the data set from the live environment for comparisons etc. until it is classified for production roll out.
Once the report has been approved and deployed to production there can be situations where some more changes have to be made on the report based on inputs from the end users. We may not be able to identify or foresee this during the development stage. These changes that are requested for are the actual feedback for the development team which indicates that their report is actually being used for gathering actionable intelligence. There may be situations where even after being deployed in production the report goes through various iterations.
The changes requested for after a production deployment can be of a different nature – ranging from user interface changes to additional queries, what-if scenarios, etc. While the reporting tool may be able to address simple requirements with ad-hoc queries, there might be some other complex requirements, like for example, the number of mobiles sold through ‘Best Buy’ in Chicago in the US during the Thanksgiving weekends for the past 3 years, analysing the trend of sale happening during a specific occasion for a specific group of customers. If the tool is not able to support such complex queries, the development team will then have to tweak the reports further and provide a solution. Therefore, until the reports are released into production and it works with real-time data, the report will not be of use.
To summarize, the following are the steps involved in delivering a perfect report:
- The Business Analyst will need to have a better understanding of the overall needs of the BI reports required by the end users
- Define a list of all the actionable insights required and also identify the what-if parameters
- The Development team should use the above as the base for developing the BI reports
- Only after the BA confirms the working of the reports to be accurate, the report should be released into production
- The BI development team should make use of real-time production data to further fine-tuning the reports
- Wait and Watch for feedback and additional requirements from end-user/real-time users
If the development team gets more requests from the end users on fine-tuning the reports, they can give themselves a pat-on-the-back and consider their solution to be a success! This also means that the end users will have gained confidence on the data that they see and will come back with more report requirements. Having said that, if the development team is not able to cater to the end user needs within two weeks, they should start working on identifying the weakness on their front and address it ASAP. It could be the choice of the tool, development/production environment issues, etc., but businesses cannot wait for a development team to take their own time to build a report. User requirements tend to keep evolving and we have to cater to them within a short span of time say, 2 weeks.
- Factors to consider for BI Reporting Tool Selection - October 17, 2014
- BI Reporting : Perfecting production ready reports - August 19, 2014