I must admit that while Charles was very excited to hear the announcement of the 20% time trial at Atlassian, I received it with mixed feelings.
The first thought that popped into my head was that we are giving people one day a week to slack off. Hence, the features that we think will benefit customers the most will be delayed. How can we encourage developers to spend time away from the chosen roadmap? Aren’t people just going to spend their time learning the latest scripting language of the day off the web? How will that bring benefit to our customers?
I could just see it…
Coming to work on a Friday morning, checking whether we are on track with the latest Jira release, and seeing the guys, beer in hand, recoding Jira’s data access layer in JavaScript!
I’ve been at Atlassian for 5 years now, working on Jira, our oldest product, first as a developer then a team lead. Over the years I have seen Jira grow in size and stability, provide more solutions and deliver more and more “bang for the buck”. It’s my responsibility to ensure that we continually improve Jira while meeting deadlines, and it’s something I truly enjoy. Therefore, I’m not a big fan of anything that threatens our ability to continue doing that well.
However, a few moments later, another thought kicked in. Choosing features has always been our hardest task. The number of possible improvements that we could make, hugely outnumbers what we can actually deliver with acceptable quality. Additionally, it is often impossible to accurately gauge the benefit of making one large improvement versus twenty five smaller ones, especially because people use Jira in so many different ways. What is extremely important to one person is not necessarily so for the next.
I believe it is even harder to prioritise feature work with respect to internal improvements. We feel guilty each time we take away developer resources from feature work today, (e.g. to speed up the build process or improve the development environment setup script), so that we can spend more time on features tomorrow.
All Atlassian developers use our products, speak with customers and do support. They are an extremely committed bunch of people, who have a keen sense of where good improvements can be made. They are also the best people to evaluate which internal improvements need immediate attention.
Rather than schedule endless discussions about relative importance of each improvement, why don’t we give developers time and ask them to use it in the best way they can come up with? I think we could see some great benefits come out of this and allow developers to identify even more with the product they are working on.
I do believe, however, that the success of this trial heavily depends on the discipline of each participant. Each developer needs to ensure they use their time effectively. I am really in favour of time boxing the effort to 20% and sticking to the guidelines that we have in place, especially the peer review of the 20% time projects.
From my perspective, the 20% time trial is an interesting take on Product Management. While the project could affect our velocity in the short term, I believe that it has a large potential to help us improve our products and deliver more customer value, via more innovative releases and internal improvements to tools and processes.
I am eager to see the results at the end of the trial.

My people are spending time on f**cking what?!