Speaking at the BBC QA Day today, Liz Keogh said that learning is often a constraint on software projects. Software projects deal with many risks and unknowns as they are mostly about new product development. Doing a project which took eight months again, with the same people and same technology, would take two months said Keogh, explaining that people would be much more efficient as they have already learned everything they need. One of the key ideas in lean software development is not to go faster than their main constraint and trying to help that constraint. Missing an opportunity to learn is waste, said Keogh, adding that focusing on learning as much as possible can help projects move much faster.
She gave an example of standups, which many teams use to coordinate work. They are a great opportunity to learn and should be used for that, according to Keogh. Instead of talking about what you’ve done yesterday, talk about what you’ve learned to make standups more efficient, she advised.
Speaking about the preferred BDD iteration cycle Plan-Do-Check-Adapt, Keogh said that the “Plan” part should be about learning how a feature should work. A lot of the “unknown unknowns” were actually known to somebody but they thought that it’s obvious and didn’t tell anyone. That’s why it is important to extract the knowledge from different stakeholders. Ask “we’re about to do this – do you know of anything that would stop us from delivering your goal if we did this”. That knowledge then becomes the foundation for BDD scenarios and acceptance tests.
In order to move faster, identify stakeholder goals that have been done before by other people or in other projects and postpone them as they aren’t risky. Get the feedback on things that are risky – and postpone things that aren’t risky – to learn important things sooner, said Keogh.
Related post: Declan Whelan: How to promote learning in software teams
|
|