I sometimes wonder if developers should do an end run around all this bullshit and come up with and start measuring management productivity metrics.
I dont see a downside to doing this.
I dont see a downside to doing this.
Let me know if you find any downsides.
Anyway, management will of course argue that developers under them are incapable of seeing everything management does. After all, management's job is to shield developers from other managers, so if the developers think all the managers at the company are worthless, that's actually a sign of how well management did their job of shielding each other's teams from each other.
These companies are even more stratified than before with the lumpenproletariat doing human-mechanized tasks while executives program their lives using software we write in exchange for the unbridled luxuries like the chance to own a roof over our head one day.
It's not exactly the future I wanted.
So, yes absolutely, administrative work now can finally be replaced, and we can free up all the tormented souls in these managerial positions to do something more meaningful with their lives.
There are also plenty of successful companies and projects without management too. Basically every 1-to-5ish-man consulting shop has zero managers and some of those do wildly well. Some of the best indie games, produced by teams of 3 or 4, had zero dedicated management. Most open source projects have effectively 0 management.
Valve, famously, kept a flat organizational structure for a long time, and certainly was somewhat successful.
What a coincidence! Thats one of the first thoughts that crept into my head when code metrics started being used on me.
This "voyage of discovery" you've alluded to is exactly what I meant by no downsides.
If managers feels threatened by being measured by their employees after their employees start measuring them, well, that's also an interesting reflection is it not?
People at the top of the hierarchy probably do care about people at the bottom, but it's difficult to care about a large number of people as individuals, so they resort to metrics that probably models what's going on. Unfortunately, the metrics don't always work.
Instead of measuring and gaming numerical metrics, enforcing a particular culture might be a better way to go:
It's the one used by all the major ones: airlines, auto manufacturers, and teachers. What institution do you know of that has a greater emphasis on measuring performance?
I'm sure most unions would love performance based pay where union reps decide what constitutes good performance. For some mysterious reason management is as keen on unions exercising their judgement as unions are on management deciding.
Where unions agree policies that treat members as interchangeable (e.g. age based pay) it is usually as a result of a compromise brooked with management who would love to have the latitude to give pay raises to scabs, kiss asses and spies.
The shape of the union is whatever the membership wants it to have.
At least programming has some verifiable realities that can be witnessed objectively by multiple observers. Not that such things are often used on "metrics", but they could be.
The quality of someone's management is hard to assess from outside, much less objectively verify. Has your manager increased or decreased your productivity today? Was it necessary that they do so for a larger goal you're not considering? Were they just power tripping?
Measuring is an aggressive act intended to invoke control that has a veneer of innocence.
That said I'm now wondering if there wouldn't be some metrics which might prove useful to workers either way.