- The lessons that I've learnt on recent Oracle BI projects
- How we could apply these lessons to develop an optimum OBIEE rollout strategy.
In no particular order, some of the key lessons that I highlighted and discussed in greater depth were:
1) Releasing too much functionality at once can overwhelm users and support processes.
2) Requirements for Reports and dashboards evolve as users gain more understanding of the tool capabilities.
3) Involve Users as early as possible in Development process.
- Especially if requirements are unclear or poorly understood.
- Report and Dashboard Prototyping often works well.
4) Ensure End Users understand OBIEE Report and Dashboard functionality and capabilities before the Design process starts e.g.
- View selectors, Column Selectors, Dashboard prompts, Delivery options, BI Publisher integration, and so on
5) Develop organization wide Dashboard and Report 'Look and Feel' standards and enforce them.
6) Review quality of data model regularly: e.g. ensure consistent mappings, conforming dimensions, and so on.
7) System performance and perceived 'acceptable' response time is key to end user acceptance.
8) Maintain a library of existing and re-usable solution.
9) End User Training should be focused and timely.
Reviewing my lessons learnt led on to a debate on iterative and agile development methodologies and how they could be best applied in an OBIEE implementation. I'll post some more on that on a future date.