Imagine being one of the first businesses to install a UNIVAC computer or roll out a barcode system for inventory management. Each action could have propelled you ahead of the market in how you recorded, processed, and used data to inform your business decisions. But if you couldn’t trust this new data you had access to, what advantage did you really gain? Today we all generate data in astounding quantities, but still wrestle with inaccurate, out-of-date, and incorrect data.
Data warehouse testing can bring clarity and confidence to BI reports so your business stops leaving money on the table. But how do you actually kickstart your BI/data warehouse testing projects? Tosca BI Architect Alex Tabisz and Tosca BI Product Manager Alain Traechslin got together for the webinar BI/Data Warehouse Testing: Why, When and How to address this exact question.
In fact, they addressed this question and many more, assembling the most common questions encountered in the field, and with Tricentis Tosca customers.
- When and where should we start automating?
- How can we ensure test coverage across different systems?
- What test types do we need to cover?
- How can we understand complex business rules and determine how much is tested?
- How do we minimize test maintenance?
- How should we optimize risks and manage the trade-off between automated and manual testing?
- What questions and challenges typically arise over the course of a BI/DWH testing initiative?
In response to the first question, Alain and Alex shared their recommendation to begin with metadata tests whenever possible, as they require very little maintenance, and answer key questions which will shape the rest of your testing investigation. For example, metadata tests let you know if your platform is stable, and which of your existing tests are affected by changes. Being able to get an immediate impact analysis is just one of the powerful insights metadata tests can uncover.
Catch their advice and tips for the rest of these questions by watching BI/Data Warehouse Testing: Why, When and How on-demand.