Done - Do you know how to make sure you deliver a build that’s tested every Sprint

Last updated by Brady Stroud [SSW] about 1 month ago.See history

What happens when you leave all the testing to the end of the Sprint? You find things that are not done and you have no time before the Sprint Review to fix them.


Figure: Bad example - if you don’t complete all the tasks the customer will not receive a build in the Sprint

One way to mitigate this is to aim for a “ test please ” to occur a few days before the end of the Sprint but you still run the risk of not having enough time to make sure everything is done.


Figure: OK example – Send the “test please” before the end of the Sprint so you have time to finish everything

It is preferable to conduct a Smoke Test to make sure that you are comfortable demoing the unit of work you just finished to the customer. One way to do this is to create a Coded UI test for each of the Stories as part of your Definition of Done (DoD) that runs through the functionality you have built.

In this scenario the “ test please ” with the customer happens outside of the current Sprint.


Figure: Good example – Create a coded UI test for each story to prove that it is complete

If you are doing Scrum then you should have a User Story fleshed out with a set of Acceptance Criteria . These Acceptance Criteria are agreed with the Product Owner before The Team committed to the story, and define what the Product Owner will accept as complete. This makes it relatively easy to create some automated tests that test for these Acceptance Criteria and help your Sprint Review go as smoothly as possible.

Any changes found during the Sprint Review get added to the Product Backlog to be prioritised by the Product Owner .

RuleBuildEverySprintUltimate Figure: Ultimate example – Create a Test (could be Coded UI) for each of the Acceptance Criteria agreed with the Product Owner

Adam Cogan
Eric Phan
Martin Hinshelwood
We open source. Powered by GitHub