I once worked at a company where the new CIO came out and said very plainly that the attrition rate was too low. The implication under the circumstances was that there were too many low competence long-timers. This was followed up with a ham handed campaign of making everybody miserable by dramatically increasing the sense of stress and "urgency" that everyone assumed was designed to increase the attrition rate. That in turn was followed by a small exodus of many of the better engineers (since that's who is able to leave easily when things go south.)
Five years later that company is now two CIOs removed from the original attrition warrior who left presumably to help the attrition rate even. The company still has a ton of long-timers that they seemingly can't get rid of. Why are companies so bad at coming up with a system that keeps more competent people and eases out less competent people but that also doesn't hurt morale by being a dystopian nightmare?
Because "easing out less competent people" a fool's errand.
There will always be high performers and low performers on a team, both in a relative and absolute sense. Good managers don't make it their responsibility to close that gap; they just make it clear what the goals are, fill the team with self-motivated people, and provide the tools and environment to get the work done. And yes, you fire people if they don't meet a clear minimum level of performance, but don't make that an expectation out of the gate. Even if you cut all the "low performers" and double the "high performers'" pay, those people can burn out or get sick and suddenly turn into low performers. "First, Break All the Rules" is a great book on this topic.
There will always be high performers and low performers on a team, both in a relative and absolute sense.
Yeah, although the more heterogeneous the tasks, the trickier it even is to define these terms.
Many years ago I was on a small (2-3 people) team with a truly genius-level coder. He was close to what you'd call a savant. He was (despite being a sweet guy at heart) incredibly difficult to work with - he was unfiltered and abrasive, and would write reams of code to solve pet problems that didn't really need solving in the context of whatever the hell it was that we were trying to do at the moment, because they were "fun" coding challenges that captured his imagination.
The end result was that while he could do the work of several engineers, you also... needed an engineer or two to work with him nearly full-time just to keep him on the rails and actually working towards the project goals, and to act as a buffer between him and people outside the team.
So who was the high performer on our team, and who was the low performer? Well, I didn't suck, but he was certainly the better engineer! But was I really a "low performer" when I did some of the equally crucial work of keeping him on the rails? Was he really the "high performer" when he required lots of people-hours from others just to keep him on task?
Those terms really lose all meaning when you're all engineers but are contributing different things to a team... the high/low performer stuff really only works when people are working in true isolation under identical circumstances. Like maybe for ditch-digging or weightlifting competitions or whatever.
It’s incredibily difficult and why managing is so difficult. Your comparison of who is the best performer is more important maybe for bonuses. This article is about firing the worst performers. So unless the process was “keep the very best single performer, fire the rest” the question of who is a better engineer is moot and doesn’t need to be solved.
It’s very difficult to evaluate articles like this without knowing the process. I could see it working if you had a simple process that relied on strong training and judgement with some governance to limit the impact of errors.
But if people get ranked based on some stupid ML algorithm and just sliced out of the company, that’s unpleasant for the fired staff, and for remaining staff. Fortunately, I think companies that do a dumb version of bottom 5% will fail or get corrected by their board and directors.
That's a great example. I think basketball, especially the NBA, is an interesting correlate for thinking about team dynamics, especially programming teams which in my experience have often been about the size of a basketball team.
I read The Jordan Rules last year about the Bulls' first championship season and before the season begins, Phil Jackson says something like, "If Michael Jordan wins the scoring title again this year, we won't win a championship." This was after they had reached Game 7 of the Eastern Conference finals the year before.
If I remember correctly, Jordan did win the scoring title that year as well as his first championship ring. But even in a sport like basketball where there are stats and increasingly sophisticated metrics and data analysis, I find it most interesting reading about the less visible contributors and the crucial roles they often play. (Breaks of the Game, the book that inspired The Jordan Rules, I think does even a better job of getting into this aspect of the team dynamic.)
This is also related to the big idea popularized by Moneyball. But where Moneyball struck me as more relevant to hiring, reading about the NBA has led me to think about how teams perform together more carefully where the metrics as you say are often trickier.
>>If Michael Jordan wins the scoring title again this year, we won't win a championship.
In the 1990s, in cricket, we would see something like this from Sachin Tendulkar. He would reach 90s, then eat 8 - 9 overs to just get his century. Net result was, he would continue to set new records for maximum centuries scored- but the team would lose.
>>This is also related to the big idea popularized by Moneyball.
It's fairly obvious to arrive to this conclusion. Having a team full of super stars doesn't win you things. Though this is counter intuitive to understand- Quality exists at various other levels apart from heroism.
Yes. There are a LOT of guys in the NBA who could score 30 points a night (good enough for the scoring title, most years) if their teams offense constantly funneled them the ball, but they'd be scoring those points due to taking ungodly numbers of shots at relatively low shooting percentages, at the expense of their teammates.
Basketball analytics have already begun to phase out guys in the mold of Kobe Bryant, Allen Iverson, or Carmelo Anthony who were great players who scored a ton but did so with (at times) low efficiency.
Ah, rockstars. If you get one without the craziness they're amazing. But the situation you've described is basically Stakhanovism with code rather than coal.
This! Which comes down to the basically stupid categories that we use to tag people (high/low performers, smart/dumb, even good/bad). Sometimes I wish for a better developed vocabulary for context, as context is much more important than any individual trait in most, if not all, cases.
I should have added "for any given metric of performance."
You have to use SOME kind of objective metric, to check your own personal biases if nothing else. But you're absolutely right in that performance metrics will never do your employees' various talents justice.
It’s easier to build somebody’s tech skills than fix their personality. That’s why team dynamics trump tech skills. If your attitude causes bad vibes in the team then you’re out.
I've worked with folks like that. They usually do more harm than good and generally need to be managed out or into roles more suitable for how they like to work.
Things worked with this guy because at his core, he was actually a very sweet person! (And we were very patient and interested in making it work, if I may say so. Not that we did everything right, not by a longshot)
He was just kind of socially clueless, as opposed to actually malignant...
> Why are companies so bad at coming up with a system that keeps more competent people and eases out less competent people but that also doesn't hurt morale by being a dystopian nightmare?
All metrics-driven systems that try to collapse complex effects to a single metric without intelligent human supervision tend to produce dystopian or at least spectacularly unintended effects.
In this case, the problem is trying to have a single axis of competence. Quite apart from the range of technologies and circumstances people might find themselves in, there are also the interpersonal skills and meta-competences that matter.
Then there's the company context: were they supported or let down? By their managers or colleagues? Were they working on something that failed for unrelated reasons? Or on something that succeeded despite their involvement?
In many ways it's the same as the hiring problem, it's just less concentrated on a single day.
Competence is a transitive property: people are competent at things, not as a global property of their personality.
> That in turn was followed by a small exodus of many of the better engineers
Engineering (especially software engineering) seems like a particularly bad field in which to have "forced attrition" or stack ranking since so much of a team's success depends on the engineers' ability to work together directly on an interpersonal level, and for their code to be easily maintained/extended/tested by others.
I suppose there are partial ways around this, such as factoring rankings from one's peers into the forced attrition.
But overall it's very difficult to imagine any engineering environment that would benefit from incentivizing engineers to compete with one another rather than cooperate.
(Lot of engineers are already good at playing the game in which they look like team players, but are really autocrats of their own little silos and knowingly make themselves hard to replace & keep others down)
These things get really weird really quick. I worked for a hardware startup that (of course) had just brutal delays in shipping their product. This included scrapping an entire product line and shipping a less ambitious, but somewhat orthogonal one.
So 3 times in two years there was a big push for the marketing, customer success, developer relations, and support etc teams to get ready for a launch that never happened.
How exactly is a customer success person supposed to be judged as a high performer if there are no customers? Support people sitting by a phone that never rings?
This is an obvious example, but I find the same thing really often in engineering groups: "I was asked to do X. I killed myself to make X happen. Turns out X wasn't relevant to the business and now it's all being shut down and I'm a lower performer."
> Why are companies so bad at coming up with a system that keeps more competent people and eases out less competent people but that also doesn't hurt morale by being a dystopian nightmare?
It's called the "dead sea effect" and it can be a tough problem to solve. I think every company has a certain percentage of low-productivity, low-competence employees that are hard to get rid of. Doubly so if the employees are unionized.
We have dead-sea effect to some extent in our company and honestly it seems our solution is to just wait until the underachievers retire and then be more rigorous and diligent when hiring their replacement.
Why are companies so bad at coming up with a system that keeps more competent people and eases out less competent people but that also doesn't hurt morale by being a dystopian nightmare?
Probably because they can't tell the good from the bad.
Nonono, that would require actually investing in your people^Wresources. You're supposed to minimize CapEx, man. Human resources are a fixed asset; depreciation is expected, and good for taxes.
Which seems like a part of the problem here. Also, since they're an expense, not an asset, why would you want to invest in them? After all, you wouldn't pay for a paint job on a car you're just leasing.
And then companies are surprised people job-hop :).
As someone who has tried that in the past: it doesn't work. I would kill myself trying to get certain people up to speed and also trying to do my own work.
The problem is that it's a self-fulfilling prophecy. For the most part, people are less competent because they don't put in the effort to get more competent. Trying to force people to give more effort is not a fruitful endeavor unfortunately.
> That in turn was followed by a small exodus of many of the better engineers (since that's who is able to leave easily when things go south.)
That is such an obvious outcome. When it seems layoff/low-morale seems to be coming, people who will leave for new jobs on their own are usually ones with the initiative, right skillset, youth, good with people, etc, which is why they can get a new job with ease. And these are the very people a company wants to hold on to. So the other option to achieve attrition rate is to have lay offs. But of course that is not good for reputation of the company.
> Five years later that company is now two CIOs removed from the original attrition warrior who left presumably to help the attrition rate even.
I have a suspicion the CIO was specifically brought in by higher ups to force higher attrition rate.
> Why are companies so bad at coming up with a system that keeps more competent people and eases out less competent people but that also doesn't hurt morale by being a dystopian nightmare?
they aren't wrong, low attrition usually means that people aren't pushing themselves out of their comfort zone... this is incredibly bad for a company (ESPECIALLY a company that uses a hierarchal structure). Loosing a little fat is a sign of good health for a company... the issue is when you stop cutting "fat" and start cutting into the meat.
Amazon is just taking advantage of all the devs that want to work for a big name, its a fair deal for both sides. They work you to death for 18 months, you get "Amazon" on your resume.
Comfort is not bad for a company period. Its what keeps the lights on, but it depends on how you define it. Most projects have horrible deadlines and they happen often so there isn't a lot of time to play R&D on a lot of things. I haven't worked for many companies that are fine with breaking what works. In fact, it often results in discipline because you were trying something that wasn't normal operating procedure mostly because the risk couldn't adequately be measured or realized (i.e. trying a new framework for example).
I for one, would not look too fondly on Amazon precisely because of the way they treat their workforce and because their product is not very good because of it.
There are systems of operating that are ridiculously profitable; I know the engineering executive of a Forge Shop that implimented a Scanlon plan some 50 years ago. Company is employee owned with the original company ownership serving as chairboard members.
The janitor, working there some 40 years, retired with $28 million in his pocket.
All it took was management and labor deciding on what it cost to run the business, and savings over that, they put some into an expansion\risk fund, and the rest they split. That technology came out of the 1930's where people were starving, there was no capital and debt to work with, riots were commonplace, and people were looking at the management as food.
Funny thing, when everyone is watching out for one another and helping eachother, really awesome results tend to happen. Sure there's a personell mix issue, but personalities can grow and change. Realistically, all corporations should be run this way. We give the executive management immunity and then we allow them to abuse the staff. That's a recipe for disaster. The majority of company stock ought to be owned by the employee's doing the work.
There's this fatalistic romanticism that goes on with being the top dog, alpha male, name your price, et-cetera. This comes from rich people who never came out of their terrible 2's. They grew up leading isolated lives and never spent time with normal human beings, so they develop this superiority complex and go on tantrums. The reason you see so much advertising that makes no sense for executives is, there are so many of these "VITO's" out there and they all sell emotionally because they are literally box-of-rocks stupid.
The funny part of the joke is people, to their own peril, give respect to these people. We've managed to dress it up with modern marketing\psychological warfare, and we've managed to let the pack of grifters get so large they're now fighting over the scraps. A City of thieves produces no bread. Fertility rates are 1.72 in the US, USSR during its collapse was 1.2. We're halfway there. We need a term to define cannibalistic management; the kind of management there that figures out ways to literally eat people by entangling them in this web of socioeconomic corporatism. We already have "trap" music so we aren't far off, culturally, from defining it conceptually.
Accountants understand risk and cost, MBA's understand Risk, Cost, and Process. It takes someone with specialized experience and skills to figure out how to build an efficient process, or to do R&D because you need to understand bottlenecks and how to cheat at gambling; effectively it takes that kind of person to understand oppertunity, of which the other 2 don't.
So you get into these ridiculous situations where an accountant or MBA is put in charge, with no real understanding of the actual systems at play and its revealed to the staff executive management is really just a bunch of permanent vacationers or grifters. The really great engineers make up excuses based on prior, painful experience as to why they should leave and do so; the existing staff figure no matter where they go, they'll have to deal with the same BS.
Accountants and MBA's always strangle companies, so if you see a new executive manager who's resume doesn't have "line" experience, it's always time to leave for greener pastures.
Your story here is why I hope to see more co-op based businesses come into being. I'm nearly the opposite of a communist, but it takes a fool to think labor shouldn't own a good amount of what they do. Particularly when your work requires pouring your soul into it.
Why fire low performers? Why not try to make them become average performer?
Maybe they have children to care for at home and doing remote work would help? Maybe they don't like some of their colleagues and a team change is in order. Maybe after months of year of being proactive but being shutdown by management they decided to stop caring? Maybe it's more serious and they should be encouraged to see a therapist.
Because management is difficult and proponents of stack ranking and other data-driven approaches think there's a way to automate management. It turns out -I believe anyways- that for many kinds of jobs you can't really automate management. If you want to know who your best and brightest are, and who the chaff is, you need to get down and dirty and get to know them.
Yeah, it's an analogue problem and not easy to solve. Hopefully, a manager that did that would get found out when his team was full of incompetent cronies.
Five years later that company is now two CIOs removed from the original attrition warrior who left presumably to help the attrition rate even. The company still has a ton of long-timers that they seemingly can't get rid of. Why are companies so bad at coming up with a system that keeps more competent people and eases out less competent people but that also doesn't hurt morale by being a dystopian nightmare?