Recently we Talked with some of our customers about QA Automation trends. We were trying to figure out the latest trends of QA Automation based on personal experience from different verticals. We came up with some important insights.
Here are a few observations you should be aware of:
1. There are still people who think about moving QA Automation for cost reduction, meaning trying to save the expense on Manual QA Engineers. While there might be a calculation that shows some financial benefits of replacing some QA Engineers, that spend most of their time walking through a long regression cycle with automated cycle, QA Automation brings other costs with it. Some examples are: more expensive engineers (programmers), a framework to develop and an ongoing test maintenance.
2. The great benefit of QA Automation is not in direct cost reduction but in moving focus from past (regression) to future (new features) where automated tests take over regression.
3. This changes the QA engineer role in the organization: from being the guy who knows the system best (a kind of a super user), that waste most of his time on endless regression cycle-to using the QA engineer “super user” expertise to work closely with product management. That is, in order to identify new feature testing scenarios, their relevancy to the system and to be one that identifies the problems she/he already knows will happen if not identified beforehand.
4. Moving to QA Automation also brings significant development velocity improvement. I would go ahead and say: when automation coverage is 80%, velocity is higher in at least x5.
5. Why is that? Developers, just like the QA engineer above split their time between focus on new feature development and making sure their work did not break anything. Actually, before QA Automation, a developer had to ask for the QA Engineer (sometimes more than one QA engineer) to feel safe before he pushed a feature to the shared repository. With QA Automation, the developer has a big muscled team of QA engineers on his side. When developing a new feature, he is able to run the full set or specific regression subsets to make sure nothing is broken. Working with an “Automated QA team”, capable of running a regression cycle that takes weeks for a manual QA team on your local or personal cloud environment within 20-120m (depending on the size of product and coverage) increases engineers confidence and enables much faster development.
6. Last but not least: the trend of moving to agile development, continues integration and the holy grail of continued deployment, has also put a lot of pressure on QA to be faster and more automatic. This makes it almost impossible to work Agile or move to Continue deployment mode without having QA automation implemented .
David Zonsheine was VP R&D for a few companies and is now Galil Software VP Sales.