How do you include manual testing in scrum?
The first question I would ask is: Why are you doing the manual testing? When we talk about modern testing techniques–things like unit testing, functional testing, integration testing, regression testing, static testing, even automated user acceptance testing based on your acceptance criteria–so many of these things can be automated. What we’d really like is for the requirements themselves to have automation testing from a system perspective. Now you’re going to use the customer representative. Under a scrum model, they’re called a product owner. Have that customer representative do the discretionary view of the requirement to make sure that it works the way that they want it to work. That’s really the best-case scenario.
What we’re really trying to avoid here is inventory. Many times what happens is people are doing manual testing and so they separate the people who are doing the coding from the people who are doing the testing. What happens is you’ll get this big inventory buildup into the very end of the sprint itself. Now you have a limited number of testers who get overwhelmed. At the end of the sprint, you’ve got a bunch of partially done work and you have very little fully done work. What we really want is to get that automated, so that you can get those requirements to complete.