A national food supplier had been gathering daily sales and inventory data from retailers for the last 20 years. Their internal architect and database team had been using a HyperRoll, which adds multidimensional query capabilities to an Oracle data warehouse, and Business Objects (Crystal Reports) to create reports that were exported to Excel.
Separate developer teams for Oracle and HyperRoll, Business Objects, and Excel VBA were all involved to create interactive reports and dashboards. Change management requirements also made manual deployments of this complex set of technologies very time-consuming and error-prone so only one or two releases were made annually.
In addition, HyperRoll became a discontinued product in the middle of this company’s technology stack. They had the warehouse already and their analysts were clear on their goals but they needed to replace HyperRoll, reduce the complexity of their reporting technology stack, and improve their velocity.
As part of Imaginet’s solution we wanted to introduce a more lean approach to rebuilding their systems and teams. Fortunately this company had used Microsoft Team Foundation Server (TFS) previously and had a history of implementing and educating staff on Lean manufacturing processes, and a desire to bring the concepts of Lean to IT development teams.
Imaginet’s BI Architect replaced HyperRoll and Business Objects with SQL Server Analysis Services (SSAS), so analysts were able to connect directly to the cubes from Excel to create tables, graphs, and dashboards.
Using an Agile Analytics approach, he coached the BI team to use Scrum as a development methodology, and TFS to track work in progress and bugs. Automation of any processes involved in testing and deployment was a high priority for the team, so an automated build and deployment process was built to bring consistency and quality to all code changes and deployments. As quality improved and deployments became consistently successful, the velocity of the team to produce business functionality increased. Users no longer needed to wait for months to get new features, as the team was able to deploy every two weeks with high confidence.
Total Duration: First value in 9 weeks