Bloggo back to the blog
Learnt from implementing automated functional testing-->
If you are like me, when scanning the posts on different forums and LinkedIn groups you’ve seen that topics about functional test automation mostly focus on different tools? Even though I represent a tool creator, I’m astonished that so few people are talking about the human aspects in implementing good automation. A tool is just one of the things that you need to consider to be successful with test automation.
Sure, automating some smoke tests and functional testing can be fairly simple, especially if you have well planned manual tests that can serve as input. However, as soon as you want to do something more with your tests, e.g. make them smarter or more complex, it takes some thinking before you start?
Since the launch of JAutomate last year we have met over 100 companies and we’ve been involved in implementing test automation at all of our early customers. We wanted to learn what mechanisms and challenges there are in order to make the implementation process as successful and cost-efficient as possible.
The experiences and conclusions are:
– Companies have unrealistic expectations on the automation process and/or the tool
– Automation is still associated with people getting laid off (even if it the need and amount of testers increases…)
– Lack of resources to properly engage and invest in the automation (catch 22…)
– Little or no planning and thinking before creating scripts
– Common to spend a great effort in creating the suite, but not enough maintaining it
– The wrong tool is chosen because project and personnel needs are not considered and because it feels safe go for the tool that can do everything rather than what’s best suited for the purpose.
Even though I made my statement in the beginning, when talking about tools – asking for one tool to cover all type of automation is like using a screw driver when you need a hammer. You might finally hammer the nail in place, but was it worth the effort and did you really achieve the result you were looking for?
One very common problem that leads to the need for automation is the shorter development cycles that companies currently face. This leads to testers scrambling to finish their tests before the next cycle or even trying to catch up to the current cycle, causing stress that leads to mistakes, poor manual or automated test cases, faulty test results and overall lower system quality. Thus, even though the testers miraculously manage to finish the test cases, there is still not enough time to plan for a future, larger and more beneficial automated test suite – there is no time to save time (this is what I mean with catch 22 above…). This is the reality we have seen and it’s easy to understand that the adoption, implementation and maintenance of any automated test tool/technique must be as quick and flexible as possible to minimize cost and raise system quality.
However, we tool creators will do our outmost to develop new functionality to help and support the tester, but in the end good quality and low cost requires human intelligence to achieve.
Author: Olle Alvemark, CEO of JAutomate
Currently living in Gothenburg but grew up in the north of Sweden. Master’s degree in automation at Chalmers University of technology. Gave up a career in energy performance contracting to pursue making Visual GUI testing and JAutomate a part of the global testing community. Feels for the human aspect of test automation and is keen on making tools fit the human and not vice versa.
Read more about their tool JAutomate at http://www.jautomate.com/
Or visit us at Stand 31 at the EuroSTAR Conference Exhibition.