Deciding what to automate – part 2

Once you what to test, find the subset of what to automate

Again everything you want to ‘test’ cannot be ‘checked’

It’s best to be smart about what should be automated

While the most common answer to this question is ‘automate what we can’

There’s a fundamental flaw in that, not necessarily all we CAN automate is WORTH automating

If a feature is hardly changed and used, automation efforts are best spent elsewhere

Conversely, if a certain feature is time consuming to test and has to be done every time, I’d rather automate that feature.

More factors IMHO to help deciding what to automate here: