Business Intelligence Cloud Implementation
From static on-prem to self-service cloud
Situation
DDI's business intelligence (BI) tech stack consisted of a few static SSRS reports, built on an on-prem SQL stack (SQL, SSAS, SSRS, SSIS).
The technology was poorly designed and implemented, resulting in a constant flow of bugs and manual updates to keep revenue reporting remotely accurate, preventing the team from putting any real time, effort or even thought into expanding or improving the platform.
To further compound the issue, the BI Product Owner had just been let go.
Task
After four months on the job leading the customer relationship management (CRM) revitalization, I was asked to take on the full job duties of the BI Product Owner, while continuing the rapid acceleration of the CRM platform.
Action
I broke the team into two focus groups, one tasked with knocking out a reduced set of production support bugs aimed at maintaining business continuity, the other aimed at rapidly identifying an MVP model of a sustainable and scalable cloud BI platform. I then augmented that vision by roping in a handful of internal and external experts, and rapidly moved into the "build" phase.
Throughout the process I also collaborated with the Finance and Oracle ERP experts within the company to gain a very detailed understanding of the data and database.
Lastly, I spent every free moment learning Power BI, in order to gain a better understanding of the realm of the possible, the speed at which progress could and should be made, and to challenge the team on designs and commitments.
Result
We delivered a 100% accurate revenue model that not only supported the existing reports, but more importantly that enabled power users and citizen developers outside of the technology department to create their own reports marrying their own datasets to the Finance-approved company revenue.
This enabled the BI Team to simultaneously cut resources and move from solely production support to platform growth and acceleration by focusing on building reusable datasets that could allow hundreds of users to create their own pixel perfect reports instead of building and supporting each report one at a time.
