I’ve talked before about the non-team team dynamic that is “one person per task”. Where the management and engineers collude to push the organisation beyond a sustainable pace by making sure that at all times, each individual is kept busy and collaboration is minimised.
I talked about the deleterious effect on collaboration, particularly that code review becomes a burden resolved with a quick “LGTM”. People quickly develop specialisations and fiefdoms: oh there’s a CUDA story, better give it to Yevgeny as he worked on the last CUDA story.
The organisation quickly adapts to this balkanisation and optimises for it. Is there a CUDA story in the next sprint? We need something for Yevgeny to do. This is Conway’s Corrolary: the most efficient way to develop software is when the structure matches the org chart.
Basically all forms of collaboration become a slog when there’s no “us” in team. Unfortunately, the contradiction at the heart of this 19th century approach to division of labour is that, when applied to knowledge work, the value to each participant of being in meetings is minimised, while the necessity for each participant to be in a meeting is maximised.
The value is minimised because each person has their personal task within their personal fiefdom to work on. Attending a meeting takes away from the individual productivity that the process is optimising for. Additionally, it increases the likelihood that the meeting content will be mostly irrelevant: why should I want to discuss backend work when Sophie takes all the backend tasks?
The meetings are necessary, though, because nobody owns the whole widget. No-one can see the impact of any workflow change, or dependency adoption, or clean up task, because nobody understands more than a 1/N part of the whole system. Every little thing needs to be run by Sophie and Yevgeny and all the others because no-one is in a position to make a decision without their input.
This might sound radically democratic, and not the sort of thing you’d expect from a business: nobody can make a decision without consulting all the workers! Power to the people! In fact it’s just entirely progress-destroying: nobody can make a decision at all until they’ve got every single person on board, and that’s so much work that a lot of decisions will be defaulted. Nothing changes.
And there’s no way within this paradigm to avoid that. Have fewer meetings, and each individual is happier because they get to maximise
progress time spent on their individual tasks. But the work will eventually grind to a halt, as the architecture reflects N different opinions, and the N! different interfaces (which have each fallen into the unowned gaps between the individual contributors) become harder to work with.
Have more meetings, and people will grumble that there are too many meetings. And that Piotr is trying to land-grab from other fiefdoms by pushing for decisions that cross into Sophie’s domain.
The answer is to reconstitute the team – preferably along self-organising principles – into a cybernetic organism that makes use of its constituent individuals as they can best be applied, but in pursuit of the team’s goals, not N individual goals. This means radical democracy for some issues, (agreed) tyranny for others, and collective ignorance of yet others.
It means in some cases giving anyone the autonomy to make some choices, but giving someone with more expertise the autonomy to override those choices. In some cases, all decisions get made locally, in others, they must be run past an agreed arbiter. In some cases, having one task per team, or even no tasks per team if the team needs to do something more important before it can take on another task.