Don’t do big-bang restructures

Reading Time: 3 minutes

Don’t use structural reorgs to chase ”transformation”.
They don’t work.
They break existing teams.
They damage morale.
They sow confusion.
And they turn one set of siloes into a different set of siloes.

Amongst the dysfunctions of transformation, restructuring is one of the worst.

Work is a complex system. It’s an organic soup of attitudes, beliefs, behaviours, mood, vision, personalities… There are no crisp inputs and outputs, just energy and activity in a network. You don’t know how to change it. Nobody knows. Stop pretending that anybody knows what an optimal org structure is until you try it. It is a patronising fallacy that anyone can know what a better structure is in advance. It’s the nature of complexity. We can only experiment in increments. Structure must be emergent not imposed.

Culture is an emergent property of the complex work system. i.e. it is an output not an input. Change the attitudes and behaviours, then that becomes culture.

What’s more, WE can’t change culture. The community has to change itself, and want to do so. We ca only change the conditions under our control and see what effect that has.
And that effect is unpredictable. NOBODY KNOWS. You have to suck it and see. Hence small steps.

It is even more enervating when the goal of the restructure is “to be more agile”, but it is done in a big bang. The irony. Big bang transformations are a big bet. Agile is supposed to be about making small bets with minimum blast radius. A restructure is never that. If you’re truly agile, you will never(? seldom?) do a big bang restructure again.

Sure they work sometimes. If a reorg is successful it’s usually dumb luck and bloody toil. The consulting firms will only tell you about their successes: even a blind squirrel finds a few nuts. A few can even be seen to have been done in a fairly successful intentional way. It takes some very good and very strong people at many levels to pull it off.

But the exceptions just prove the rule. In fact I doubt there are exceptions that truly “succeed”. I’m sure it is too narrow a definition of success. Reorgs are violence. They’re done to people against their will by force. They have all sorts of toxic cultural repercussions. This at the heart is my deepest objection to how reorganisation is done. What are the medium term consequences for psychological safety, for trust in management, and for real lasting change, when process reengineering and organisational restructure is designed behind closed doors then imposed by decree?

Organisational structure should emerge from the needs of the people, pulled by them not pushed on them. It should happen in an agile way: constantly reflecting, adjusting, improving. Iterate, increment, experiment, explore. I wrote about this in a DevOps context 3 years ago, but it’s everywhere, and it’s firing me up.

You can’t make knowledge workers do anything. We have to stop doing change to people. If they don’t want it then that’s the problem to be addressed. You can’t force them to believe something different, or even to act in a different way. You can force slaves, manual workers, and clerical workers because you can see the output per person. Taylorism. Knowledge workers work collaboratively on invisible work.

And they sure as hell won’t change in a matter of weeks or months. Humans aren’t machines. We aren’t Human Resources to be manipulated and engineered.

Stop it. Be human.


This article was originally posted on the IT Skeptic blog