About the customer –
A leading telecoms provider in Germany providing mobile and fixed line services to the ‘smart buyer’ segment. This customer has consistently been voted best operator in Germany in their service segment and are widely acknowledged for their highly innovative methods in software delivery.
Following the launch of a new IT stack, the customer was quickly able to take the lead in it’s target segment in the German mobile market. Facing fierce competition from rising competitors, the imperative to defend the market position with better tuned service offerings became urgent. It was required to unlock the potential of aggregated customer data by implementing a professional business intelligence framework. The challenges were numerous:
- multiple redundant systems and data sources
- legacy systems not able to sufficiently scale to cope with growing data volumes
- home grown reporting systems
- in many cases poor data quality
Establish a single source of truth: Integrate data from various sources (CRM, Billing, Suppliers e.g. logistic, sales partner) and decommission home grown reporting and analysis systems Gain new insights: Validate financial data to discover revenue leaks, track the success of business development measures (such as advertising campaigns, bundled service offerings, value added services etc.), and understand customer structure in order to design and build new service offerings. Empowerment: Give business users the tools to build and perform real time data analysis without needing support from technical staff, enable faster decision making, enable realtime measurement of business critical KPI’s (e.g. ARPU), evaluate the potential of new business services on the basis of real production data (simulation).
For this project we used an agile method called scrum to deliver the software fast and with sustainable high quality.
Working with continuous integration and fully automated testing, we were constantly able to give a clear project and software quality status. The detailed stages of the project are described below. But note… stages were not run in sequence but mostly in parallel. The first reports and business analysis was already possible after just a few weeks!
Stage zero – Establish a product vision that acts as the project’s true north
Stage one – Design a robust and scalable BI strategy and architecture. At this stage, an abstract design of the target architecture was in focus.
Stage two – Tool evaluation (Vendor selection). Selecting and validating tools to meet business requirements. Close interaction with business users during the selection process.
Stage three – Build the first product version that is the basis of subsequent developments and then set up the product backlog to move forward with a truly agile approach. Fine tune requirements and finalise tool selection. This stage involved intense interaction with business units; ensuring that their current requirements were well understood and also advising them on future potential application of the planned technology.
When working with SCRUM, it is important to have a product version the customer can touch and work with very quickly. Every two weeks, a new version with more features is added. The direction of development can change on short notice and due to automatic testing, there are no issues with quality.
Stage four – Data quality & data migration. In this stage, the challenge was to perform the initial data load from multiple source systems while in parallel identifying and correcting data quality issues.
Stage five – Implementation of Backend & Frontend Services. In parallel, the services and software needed to constantly load new data were implemented. Appropriate technology was used to fully automate the Extraction Transformation and Loading (ETL) processes.
Stage six – Quality assurance. Throughout the process, we never lost sight of our agile delivery principles. Automated tests were built for every module and unit of code that was delivered. In parallel to the development process. By sticking to this agile continuous integration model, the entire software sweet could be regression tested automatically in less that one hour.
Stage seven – Support IT Operation. After successful implementation, we continue to support ongoing operations and provide training to business users.
In this particular case, ORACLE 11g with Informatica Powercenter 9.1 and SAP Business Objects 4.0 formed the core of the implemented system. However, we are not affiliated with ANY system house and ALWAYS choose the solution that is best for the customer.