3 ms·

I read a lot of criticism of Daniel's approach in this discussion. People don't like the fact that he wrote an app and tried to deal with the problem objective

I read a lot of criticism of Daniel's approach in this discussion. People don't like the fact that he wrote an app and tried to deal with the problem objectively. But, really, how is this so different to any other objective approach? Here in the UK there is a significant body of legislation around the process of making people redundant, so it would have worked differently, and particularly it will work differently for larger numbers of redundancies. There is always some sort of objective scoring criteria. For larger redundancies these are agreed with employee reps as part of the consultation period. You might be able to get away with less rigour if you're only making a handful of people redundant. Regardless, those scoring criteria and the scores associated with them are going to end up in some sort of model, which might be as simple as an Excel spreadsheet, but there's nothing to stop you implementing it in code. The model is generally published as part of a consultation pack so that staff can understand how they are being evaluated, and provide feedback via their employee rep. The end result is that the people sorted at the bottom, however many you need, are made redundant. It isn't nice. But it is objective, and it's important that it be so, because that's as close as you can get to any sense of "fairness" - not that it's going to feel fair to the people losing their jobs. You have to be able to justify these decisions, to demonstrate that you've gone through a fair and objective process to arrive at them, because you might land in court if somebody challenges your decisions. It's not nice, and it's always going to seem heartless to some people. But that's part of leadership: sooner or later you are going to have to make decisions that some (or even many) people don't like, and that you don't like either. In many of these cases there will be no perfect choice and no perfect outcome. There is no avoiding it. If you aren't willing to deal with this, don't become a leader.

troelsSteegin 6y ago
I would make a distinction between "objective" and "explicit". Daniel's model is product of his judgment. It's his perogative and responsibility as a leader, and he's well intentioned, etc, but his judgment is subjective. I am not familiar with the UK system but providing explicit criteria and an oppty for feedback seems like a step forward relative to the US. It's a question as to what else would need to change in US industry for that approach to play well. Pooling diverse judgments is a way to produce more robust predictions. An explicit model can help multiple people engage. Daniel did not mention discussing the re-designed org with any of his team. A less top down approach than his might have included seeking feedback about direction and team composition from members of the team itself.
fpig 6y ago
> But it is objective, and it's important that it be so But it's not objective at all. He still subjectively evaluated each individual employee, and fed this into the program as an input, in order to help him find configurations that fulfill certain conditions. And this is a good thing. I'm not 100% sure what you mean by: > There is always some sort of objective scoring criteria. For larger redundancies these are agreed with employee reps as part of the consultation period. ...but if I am guessing correctly, then this is fortunately NOT what he did here. Any objective metric in a field where there's no clear way to measure performance will always be extremely poor, despite employee reps' / unions' love for them. What would this metric be? I can remember one example (not in CS but in another hard-to-measure-output field), where it is just how long you've worked there. So if there are 5 people doing job X and 2 need to be fired, the newest 2 get fired. I'm entirely confident in a former team lead, now CTO's ability to beat the accuracy of that metric with his subjective evaluation of each individual's abilities (regardless of whether he uses a program to help him optimize based on that information). The problem people have with this is that they think the person making the subjective decision might have some kind of bias. And some bias definitely exists, pretty much 100% of the time. But in reality, even a biased boss who's worked with those people for a while is likely to evaluate SD employees more accurately than some arbitrary metric. If you don't deserve to be fired, but get fired anyway, it's equally unfair whether it was because the boss really likes a shitty employee who should have been fired instead, or because the shitty employee has more tenure at the company and they're firing the newest people, or because they threw a die to pick whom to fire.
valuearb 6y ago
There were 17 employees. You don’t need a program to solve that. And his first step should have been to the CEOs office to find out why this clusterfuck was thrust on him out of the blue. He’s showing he’s not a leader, just a Jr Manager given a fancy title.

troelsSteegin 6y ago

I would make a distinction between "objective" and "explicit". Daniel's model is product of his judgment. It's his perogative and responsibility as a leader, and he's well intentioned, etc, but his judgment is subjective. I am not familiar with the UK system but providing explicit criteria and an oppty for feedback seems like a step forward relative to the US. It's a question as to what else would need to change in US industry for that approach to play well. Pooling diverse judgments is a way to produce more robust predictions. An explicit model can help multiple people engage. Daniel did not mention discussing the re-designed org with any of his team. A less top down approach than his might have included seeking feedback about direction and team composition from members of the team itself.

fpig 6y ago

> But it is objective, and it's important that it be so But it's not objective at all. He still subjectively evaluated each individual employee, and fed this into the program as an input, in order to help him find configurations that fulfill certain conditions. And this is a good thing. I'm not 100% sure what you mean by: > There is always some sort of objective scoring criteria. For larger redundancies these are agreed with employee reps as part of the consultation period. ...but if I am guessing correctly, then this is fortunately NOT what he did here. Any objective metric in a field where there's no clear way to measure performance will always be extremely poor, despite employee reps' / unions' love for them. What would this metric be? I can remember one example (not in CS but in another hard-to-measure-output field), where it is just how long you've worked there. So if there are 5 people doing job X and 2 need to be fired, the newest 2 get fired. I'm entirely confident in a former team lead, now CTO's ability to beat the accuracy of that metric with his subjective evaluation of each individual's abilities (regardless of whether he uses a program to help him optimize based on that information). The problem people have with this is that they think the person making the subjective decision might have some kind of bias. And some bias definitely exists, pretty much 100% of the time. But in reality, even a biased boss who's worked with those people for a while is likely to evaluate SD employees more accurately than some arbitrary metric. If you don't deserve to be fired, but get fired anyway, it's equally unfair whether it was because the boss really likes a shitty employee who should have been fired instead, or because the shitty employee has more tenure at the company and they're firing the newest people, or because they threw a die to pick whom to fire.

valuearb 6y ago

There were 17 employees. You don’t need a program to solve that. And his first step should have been to the CEOs office to find out why this clusterfuck was thrust on him out of the blue. He’s showing he’s not a leader, just a Jr Manager given a fancy title.