After a month of automated test development, I was asked by the product managers how many tests had been automated. I replied that only 3 tests had been completed, which they did not find that impressive. The software is used to control a scientific instrument that is designed to run measurements lasting anything from 10 minutes to several days. These 3 tests were covered functionality that required running measurements lasting several hours. By automating these tests I had significantly reduced the time required to test the software. This information impressed management a lot more.
Test automation takes significantly more time to prepare than manual test cases and as a result there can be a significant amount of time before any value is gained from it. Proving the value of automated tests can be significantly more difficult than actually developing the tests in the first place. So how do we maintain confidence in the tests?
In this talk I will discuss how I managed to get value out of test automation early on in the development by identifying a problem which we wanted to solve using automated tests and then choosing tests to automate with the aim of solving this problem. Teams that automate for the sake of it will not solve any problem which means that there cannot be any value from these automated tests. Value can be demonstrated by showing how this problem has been solved.