It's interesting. Many large companies routinely fire the bottom 10% of their employees every year. I wonder if they would perform better if they routinely fired the top 10% of their employees (by org chart, not by performance ratings) and let talented new blood bubble upward. The Peter Principle says that people rise to the level of their incompetence, and in general new employees enter at the bottom of the organization, so logically you would have more incompetent people at the top than at the bottom.
There're a bunch of organizational anti-patterns that could be avoided with this scheme, too. Empire building would be disincentivized because it'd always become someone else's empire after a short period of time, and organizational politics is reduced as the players keep leaving, and you'd have to think in terms of building institutional knowledge from the beginning of your tenure unless you want everything you've accomplished to be undone by the next guy.
The "fire the bottom 10%" thing is called stack ranking, and it has been studied over and over and it never leads to good results. At this point I don't think that many companies are still doing it- the biggest example, Microsoft, killed it off in 2013.
Large orgs regularly do this. E.g., most of the banking industry, esp. the non-wall st side, have had layoffs every few years for awhile now. It can be called all sorts of things, and it adds up.
When you think about it, it makes a sad sort of sense at scale. Hiring mistakes mean, at scale, a good number of jobs aren't fits for the people you hired, half the people out there are below average to beginwith, and a good number end up toxic (they slow down orgs!), and over time, these folks cost more money and suborgs decay. We've all inherited WTF+NIH projects and worked with people who drain energy. So even if an org doesn't do explicit layoffs or stack ranking, and hiring committees and employees have good intentions, a healthy org should be fixing hiring mistakes somehow: it can't all be retraining / reshuffling the deck. Arguably, the numbers mean most tech companies are _unhealthy_ orgs because the high demand for engineers makes the healthy level of rehiring tough to do.
The result is toxic devs get retained and the tail wags the dog from a business perspective. E.g., I bet some of the #MeToo issues in tech, deep down, relate to the difficulty of curating a top org at scale, with very few exceptions (Netflix?)
At Microsoft itself. If a manager has $100 budget for hikes and rewards, she has to invisibly stack rank to distribute it with some logic (and not give $10 to each in 10 people team)[0].
Stack ranking was not about firing at Microsoft (or any other place), it was about distribution of bonus and hikes.
It’s called a corporate restructuring. Or they’ll just fire a few people here and there until they get to 10%. They call it “we re heading in a new direction”
> I wonder if they would perform better if they routinely fired the top 10% of their employees (by org chart, not by performance ratings) and let talented new blood bubble upward.
IIRC, the US military follows something like that practice. It's called "up or out":
> ...the 1980 Defense Officer Personnel Management Act mandates that officers passed over twice for promotion are required to be discharged from the military.
IIRC, the idea is to prevent people who lack greater potential from hogging the intermediate positions that others need to advance.
I think it's slightly different in that it would only apply to the top 10% of the org chart. A first-level manager passed over for promotion wouldn't be dismissed; only VP or C-level executives would, and they would be regardless of whether they were recently promoted. (In fact, because the positions are periodically vacated, they are guaranteed to have been recently promoted.)
It's much closer to term limits and the elected-office/civil-servant split in a democratic government. That has its own set of problems, but is generally fairly good at discouraging empire-building. It also differs in that civil servants generally can't transition to being elected officials (despite their qualifications, they face a lot of obstacles to winning elections), while here executives would generally be drawn from the ranks of ordinary management, keeping the ranks under them dynamic.
Why not apply this to first-level managers, too? I've seen more than one first-level manager who wanted to build an empire. I think all of the issues mentioned here about VPs also apply to managers.
Why would anyone move to management then? At Google a lot of L6 swes are put in the position of TL/M due to their ability to guide technically. This would probably vanish.
> A first-level manager passed over for promotion wouldn't be dismissed
This system will encourage first-level managers to do everything in their power to be "passed over" every single time. I can imagine the ridiculous shenanigans they'd pull, walking the fine line between not being too competent lest you get promoted, without being too incompetent that you get fired outright. Sounds like the making of a truly middling culture: questionably competent-ish, but not ambitious.
I think I'd watch a Office Space/Silicon-Valley-type show based on this premise. The overachiever character perennially delegated to bug triage, doing endless interviews for a perpetually open position on the team and being sabotaged when they manage to put some real work in after-hours (for comic relief. In real life, they'd get fired)
Why would you prefer to do this rather than not up-and-out. Your original reasoning seems well-met by up-and-out. Fresh blood comes in, Peter-principle people get kicked out. Looks pretty good.
When you reach term limits, you don't get kicked out of the country. It would create a powerful disincentive for people to move up at all, unless they wanted to leave the company soon anyway.
> the US military follows something like that practice. It's called "up or out"
During my time as a DoD contractor, I saw this happen.
One particular case was a mid-range officer [I'm never good with ranks so don't bother asking] who was up for promotion; it was an open secret that if he didn't get it he'd retire into civilian [and presumably commercial//industry] life. He didn't get the promotion, retired from military, and was duly replaced.
It would have the same effect that culling the bottom 10% would have. You promote mediocrity. Think of it like this: employees game the system by avoiding the "fireable" qualities and rush to the "hireable" qualities.
So everyone in the org wants to appear like a young firebrand and hide in the school of fish. As outliers are singled out, those who blend in well and play politically with their managers have the highest survival rate despite not being the most productive, as they devote most of their energy to not standing out.
In any persistent environment when you add pressure, you develop a counter-reaction. You can't change human behavior with simple levers, but you can bring out very human behavior by doing so.
Good managers are extremely valuable, and new blood is not necessarily talented.
Maybe a solution could be to make time-limited leadership positions. After you finish your time, you get another different position but you stay in the company, with the same or higher pay, so that it doesn't feel like a demotion. Permanent leadership positions will be given to those who did their best during their temporary stay. As a bonus, because these managers tried out different jobs, they will have a broader understanding of what people under them do.
You'd have to encode it in the bylaws at the time of incorporation, and then enforce it upon pain of dissolution of the company (so that if a rogue CEO decides they want to stay after all, the rest of the world is like "But now there is no company. Have fun with your paper throne.")
It's a lot like how no ruler will ever abdicate and still remain ruler, so for millenia there was no way to remove a king short of death or coup. We got around that with the invention of the constitution and the definition of a nation-state in terms of that piece of paper, which states that the ruler is defined to have set terms and be subject to an election every so often, and if they don't you get undefined behavior.
They aren't as powerful as you think- it's not like they get a huge say in how things are run. If the board was a reorg they have the power to put the people in place to make it happen, regardless of what some random Director or VP has to say about it.
It's pretty common for C-level executives, and CEO's in particular, to be pretty chummy with the board of directors. They're all in the same social circles.
Often the CEO of company A will be on the board of company B, and and the CEO of company B is on the board of company A, so there may even be a quid-pro-quo element.
While the CEO officially answers to the BoD, in reality there's a lot of "understandings" and "gentleman's agreements" going on. This is why under-performing CEOs seem to stick around so long.
Yeah, but that's a world of difference from "director level and above", particularly at a company like Google where there are literally hundreds of VPs and Directors (if not a couple thousand).
I saw the link here on HN first. Seek out the Gervais Principle. It is a more interesting view of a person’s place in a company. I have seen the Peter principle in action and have myself benefited/detrimented from it. But the Gervais principle is a whole ‘nother level.
There're a bunch of organizational anti-patterns that could be avoided with this scheme, too. Empire building would be disincentivized because it'd always become someone else's empire after a short period of time, and organizational politics is reduced as the players keep leaving, and you'd have to think in terms of building institutional knowledge from the beginning of your tenure unless you want everything you've accomplished to be undone by the next guy.