with Robot Framework
At Dutch Lotteries (Nederlandse Loterij), we work in a single Robot Framework repository with 14 testers across 9 teams. Due to the number of testers, we encountered many issues. Each of the 14 test suites used a fundamentally different approach. The same functionality was built everywhere but never in the same way. Even the most experienced programmer could not read the code.
We archived multiple particularly awful test suites. They got so bad that we gave up so we could start from scratch. Some were archived twice! We didn't see a green pipeline for months
Building quality test automation at scale is hard, especially when people of various skill levels are involved. This results in clashing approaches, accidentally rebuilding existing functionality, bad code, and inconsistency. We'll take a look at what worked for us:
-
Test Automation Guild to raise skill levels
-
Required reviews as a teaching tool
-
Build well-documented and flexible shared functionality
-
Structure the repository for clear scopes and flexibility
-
Write team-specific standards
-
Make lint rules to enforce our standards
These approaches will also work for you. Prove me wrong.
