Salaried developer employees are different from founders or executives. I can measure a founder's performance in terms of number of users or amount of revenue. I've yet to see a meaningful numerical measure of developer output. It is possible to meaningfully grade developer output only in qualitative terms, not quantitative ones. Quantitative measures are oppressive. If you want to rate me, I want someone to pay attention to what I'm actually doing, and I want someone who is capable of understanding what's going on. That's not asking too much. But that kind of investment means one manager is capable of managing fewer people, and you have to have really talented managers, and so people don't like it.
I have no problem with accountability. Very simply, if you need to keep tabs on me (already a step down in the relationship), then I expect you to do it in a qualitative fashion, with a person capable of evaluating progress. Closed tickets and hour estimates and burndown charts and all the rest of that stuff is a fiction, and it makes an incompetent manager look real busy when he is actually doing nothing. And it wastes my time when I'm trying to do important work.
They call these metrics time saving devices, but if you don't have time to go around and peer at what's actually happening in the shop, then you aren't actually evaluating anything.
You don't need a person to track metrics. If these fictions worked you could set up a machine to track progress and send reminder emails, and even to fire people who were consistently too far behind.
To put it in a not nice way, the managers at startups I've met were experienced people who weren't skilled enough to cash out during the first bubble, or who never made it to an executive position or to starting their own company. They were untalented, and their advice was bad. If they had been genuinely capable of showing me how to perform 10% better, I would have received them as saints, or as angels from on high, and would probably have spent as much time around them as possible. That was not the case.
The other reason people like metrics is because of the squeeze. You're already operating under the assumption that I might not be working at full potential, that's why you want the additional accountability. If I am working at full potential, and you assume you can squeeze 5% more out of me, and start to press down, then maybe you do get 5% more in the short run, but in the long run you've caused a problem.
In their less guarded moments I've heard people in these positions claim their job is to improve performance by just a few percentage points, and that will justify their position. Multiplied over a department, these are huge gains. Unfortunately, that's an exceedingly difficult proposition for someone without special insight, and the only way it's going to happen is if they are in control of the results metrics. Which they are.
Maybe you can recover a few percentage points in a clunky, inefficient company. In a hot startup one overseer is not going to squeeze that much out. They're just annoying.
I doubt you can get rid of middle managers. They're there for a reason. Someone has to digest what's going on for the higher ups. But there's probably a better way for them to do their jobs than inserting themselves where they don't belong.
I have no problem with accountability. Very simply, if you need to keep tabs on me (already a step down in the relationship), then I expect you to do it in a qualitative fashion, with a person capable of evaluating progress. Closed tickets and hour estimates and burndown charts and all the rest of that stuff is a fiction, and it makes an incompetent manager look real busy when he is actually doing nothing. And it wastes my time when I'm trying to do important work.
They call these metrics time saving devices, but if you don't have time to go around and peer at what's actually happening in the shop, then you aren't actually evaluating anything.
You don't need a person to track metrics. If these fictions worked you could set up a machine to track progress and send reminder emails, and even to fire people who were consistently too far behind.
To put it in a not nice way, the managers at startups I've met were experienced people who weren't skilled enough to cash out during the first bubble, or who never made it to an executive position or to starting their own company. They were untalented, and their advice was bad. If they had been genuinely capable of showing me how to perform 10% better, I would have received them as saints, or as angels from on high, and would probably have spent as much time around them as possible. That was not the case.
The other reason people like metrics is because of the squeeze. You're already operating under the assumption that I might not be working at full potential, that's why you want the additional accountability. If I am working at full potential, and you assume you can squeeze 5% more out of me, and start to press down, then maybe you do get 5% more in the short run, but in the long run you've caused a problem.
In their less guarded moments I've heard people in these positions claim their job is to improve performance by just a few percentage points, and that will justify their position. Multiplied over a department, these are huge gains. Unfortunately, that's an exceedingly difficult proposition for someone without special insight, and the only way it's going to happen is if they are in control of the results metrics. Which they are.
Maybe you can recover a few percentage points in a clunky, inefficient company. In a hot startup one overseer is not going to squeeze that much out. They're just annoying.
I doubt you can get rid of middle managers. They're there for a reason. Someone has to digest what's going on for the higher ups. But there's probably a better way for them to do their jobs than inserting themselves where they don't belong.