A single case for which deadlines are good

Deadlines are shit. In software development, at least. Deadlines are normally established based on estimates. But estimates are mostly wrong, if not always. They also can be established based on arbitrary circumstances, unrelated to the work that needs to be done. In which case there are also estimates, but heavily conditioned. That is, very wrong estimates. Working with deadlines usually make me feel anxious. And they can cause that bad kind of stress that will make you fall ill. They’re shit.

There’s a principle called “fixed time, variable scope”, that I know from the Shape Up book by Ryan Singer. A deadline is also needed, but only to create a sensible “tension between time, quality, and scope”. So it’s more like a plain end date. This is good and again, needed. The same way any event with a fixed duration comes to an end on that date you knew beforehand. Be it the Olympics, a two-day conference, or your holidays. In a “fixed time, variable scope” project, you also won’t be adding up more work in the process, as you would with any SCRUM-ish system, rather the opposite.

However, the actual one case where deadlines are useful is when asking for non-urgent feedback to a group of people. In a sane and well-defined async workplace, most times you won’t call a meeting or start pinging people with direct messages. You will write down that thing you need from them, give it just enough detail and context and post it in the right place at a reasonable time. Remember you want to make it easy for them to answer. Then you can kindly add a deadline: “I need this before Friday”, for instance. Give them at least a day. Until then you can focus on something else. By Friday, or earlier, you’ll have enough feedback. And you move on.


Thanks to José Ardila for proofreading and feedback.

Share on Twitter / Share on Hacker News