Why Are So Many Zappos Employees Leaving?
theatlantic.com
theatlantic.com
Extrapolating a general condemnation of an entire business system like Holocracy from a single data point is bad journalism, to say the least. Casually lumping all "flat" structures together and cherry-picking research to support a hypothesis (hierarchies work!) is lazy, and does the reader a disservice.
I'd be very interested in some actual reporting to see how things are going at Zappos. It's possible that Holocracy is not turning out well. It's also possible that the turnover was exactly what they needed, and a positive new model of work is emerging. Unfortunately, this article didn't do any reporting, so it's unable to serve as anything more than a traffic-generating question mark.
Good managers are glorified butlers. Their job is to do everything and anything they can to accommodate those whom they manage, and recognize that because they aren't actually creating anything, they defer to those who can and do whatever they can to help them do it. If they came from a creator position, then their main challenge is to not forget what it was like to be a creator and advocate for them as though they still are.
One of the major problems with the whole concept of managers is that they are typically higher than the creators in the corporate hierarchy, and often get corrupted by that. They are proud to be managers because it's a promotion in most organizations, but they have to deal with the uncomfortable fact that they no longer actually make anything (or never did to begin with). Thus, they try to find other ways to make themselves feel valuable, and that often involves inventing extra work for the creators to do.
If I ran a company, managers would be butlers. Their compensation would be directly tied to the level of happiness of the creators they are serving. I might even let the creators they serve set their salaries. I currently have a manager whose salary I would set higher than my own because he completely embodies this type of management style. If anybody on the team needs anything, he does it with a smile. He is our advocate, our cheerleader, our mediator, our friend, and our faithful servant. It's not a coincidence he's gotten employee of the year (based on embodying corporate values) two years in a row.
I really appreciate what Zappos is doing here and I sincerely hope it works out. The worst thing that could happen is they lose their resolve and give up on it. That would be a very unfortunate signal to the rest of corporate America that the status quo is working just fine, and it would make me very sad indeed.
I've experienced first hand what it's like to work for someone you could call a "servant leader" and also someone who's autocratic to the core.
"Situational contexts are not considered."
Being able to effectively self-manage under a servant leader requires a certain level of maturity and confidence that isn't always present in people. Actually being able to bring these qualities out in people who don't naturally exhibit them is a real challenge, and depending on the individual, sadly, can be very difficult to achieve.
If you're lucky enough to work under someone who's responsibility it is to make you better (i.e. a wise and experienced mentor), make sure you grab the opportunity with both hands and learn as much as you can in the time you have together.
I am willing to bet there is correlation between the two, but not enough to make it as a primary metric.
I think this is one of those cases where until we know with confidence one way or the other, we should probably just start out with the assumption that worker happiness is a good thing and put the burden of proving otherwise on the scientific community. Until then, let's go with our gut.
I know plenty of creators that don't want to spend their days blocking and tackling. Most "creators" just want to sit in front of their computers all day getting stuff done. With Holocracy those creators you talk about would assume more responsibilities and can no longer hide behind the servant.
Lastly, why are middle managers always seen in such a poor light? Why isn't a CEO/CFO/CIO seen in a lesser light since they really don't produce anything.
And yes, I used this post as an opportunity to rant about managers in a vertically structured company, but I see no reason why the type of manager I prefer could not exist at Zappos. In fact, it seems like it would be more likely to occur, given that they would no longer outrank the creators.
The purpose of the manager is not to squeeze productivity out of a developer as you might blood out of a stone (although some seem to think it is), but to try to make sure that the developers are working on the right stuff for the company. In a healthy organisation, that can be done as part of a dialogue with the developers and takes the form primarily of communication.
In my experience that responsibility has fallen on the CEO or the Product people.
For a good developer "what they want to build" in the context of their job is at least somewhat informed by what they think will be valuable for their company. Now, it's completely true that developers might be prone to particular kinds of getting it wrong (let's build a framework!), but that's true for any kind of person including managers.
Point being, managers are there to help make sure that the producers are informed and aligned about the wider context.
Consider Apple. Apple had a series of CEOs who were unable to stem the slide into bankruptcy. Enter Jobs, who turns a company 90 days from bankruptcy in the biggest corporation in the world. He did this with the same group of employees.
The only thing that changed at Apple was the CEO.
Consider how Microsoft has rather dramatically changed course with the ascension of Nadella, with an equally dramatic effect on the stock price. This is very clearly a result of the CEO changing direction.
Company culture as well flows down from the top, this becomes clear if you've worked for different corporations.
If the CEO doesn't matter, I wonder who would risk $billions in corporate valuation installing some random guy off the street rather than a carefully selected person.
Their function vis-a-vis you is almost inconsequential. They exist so that those above them can direct and receive feedback from you (or more accurately, the 10s/100s/1000s of you) in an efficient manner. Whether you think that's necessary or not is up for debate, but examining the manager's role solely from the perspective of an employee is only half-informed.
Personally, having to individually convince every software engineer in a company of a particular course of action sounds about as much fun as counting grains of sand.
But how do you then rectify the power dynamic between the presenter of stakeholders' views/requestions (i.e. the manager-servant) and the person with the authority to alter workflow (not the manager-servant)?
It seems ill conceived to filter communication through a medium that can effectively be ignored.
If you apply processes that were designed for average or lowly motivated teams to highly self-motivated teams, you lose efficiency. But if you apply processes for highly self-motivated teams to other teams, equally bad things happen.
So the optimal solution is that employees need to be managed according to their individual needs. Some people need managers. Some people hate managers. If a company really wants efficiency, then they'll recognize both of these and cater to them differently.
Unfortunately, I think in the the real world the highly self-motivated individuals just get recognized as "talented" and pushed up into management themselves. Which is probably about the most terrible outcome. :\
My experiences might be unique, but I've never been part of a team that at any point preferred taking longer to finish something than necessary. Creators' pride is very much wrapped in how quickly they can deliver creations. So much so that it can be devastating to moral for a team of creators to miss a deadline. It's viewed at as a failure and a sign of incompetence, even if there were valid reasons for it happening.
I think the important distinction to be made is the type of creation that's happening. If you're an assembly line worker or a subway sandwich maker, your self-identity probably isn't tied to your job. If it takes you longer to make a sandwich or a widget than someone else but your compensation isn't affected, then you probably will take longer because it is, like you said, "easier".
If, however, your identity is very intimately tied to your creations, then you will care deeply about their quality and the efficiency with which you can create them. Perhaps the main problem is managers creating creators whose identities are tied to their creations like creators whose identities are detached from their creations. And with that distinction, I would agree with your statement that "employees need to be managed according to their individual needs."
I guess because we were talking about Zappos where I think I can safely assume the creators' identities are tied to their creations, which is why I wrote my original comment from that perspective.
> Unfortunately, I think in the the real world the highly self-motivated individuals just get recognized as "talented" and pushed up into management themselves. Which is probably about the most terrible outcome. :\
Absolutely, and I've seen it happen time and time again. Often those types of people aren't very good at managing people.
If you work better happy... fine, that's a consideration. But really it's the only one. The goal isn't for you to be happy. The goal is to get tasks accomplished. One keyboard breaks you go get another one. The store shelves are full of them. And that's how the pyramids got built.
Not only do I work better happy (I do), I stick around longer. And that's really the key, here. Sure you can squeeze productivity out of creators while disregarding their happiness, but everyone has their limits. They will eventually leave and take with them a wealth of tribal knowledge. Corporations are notoriously bad at measuring the cost of losing employees with domain specific knowledge about their products or services. The attitude is typically that creators are interchangeable like cogs in a machine.
I don't know, I think there are chronic reposting lately in all categories, just not this topic. Don't we have a bot to remove duplicate posts? Because it doesn't seem to be working these days.
Let's take the simplest case of this.
You're a one man business, developing software perhaps. You're doing well, you're successful and you take on someone to help you out.
Are you now their butler? Does that person set your salary?
If not, at what size of organisation does manager-as-butler kick in?
I also think the divide people are commenting about between manager and creator is a red flag for a bad organization. If you are managing programers you should still be writing code a few layers up (same for other creators or operators). The “I’m a people person” argument doesn’t hold by itself. The Commanding Officers of the fighter squadrons I served in flew a lot and also managed 300-700 people. Certainly managers (or ‘leaders’) need to spend time on higher level issues and looking after the team, but they should still be able to operate / create enough to understand what is going on and lead people in the execution of the job. Sure, when the captain or admiral went flying it was more of a token gesture, but when you have a carrier and 5k people to look after or an entire battle group, well at that point keeping all the balls moving is challenging enough.
But a small to mid size team—if you can’t operate why should operators follow you? While certain management skills are transferable, just because you lead a team of programmers doesn’t mean you can lead a team of SEALs or vice versa. Operational experience and ability to execute are important and an MBA isn’t a substitute. Presidents of sales are still involved in selling. Why is it on the technical side where people seem to either stop executing and manage or just be a ‘creator’. Like I said, now that I’m a programmer, this is a red flag for me. Although, to be fair so is the other extreme, the super coding leader with 40 people reporting to him and no time to talk to anyone. One-on-ones every other year! Leadership by negligence.
There's some truth to that at every company, but ultimately money and ownership matters. And it's delusional to think that employees do, or should, feel unbridled affection for their employers. The only way that ever makes any sense in big companies is when every employee is not only hugely empowered but also has a substantially stake in the company.
Using "revolutionary" new forms of management like "flat" or "holocracy" or what-have-you aren't going to magically transform a company into something entirely different. At the end of the day there are people who can fire anyone and people who can't, there are people who live in mansions and drive luxury cars and people who don't. The only way to have an actual co-op is to, you guessed it, legally be a co-op. You can't just play make pretend.
In the end, to be honest, I'd rather have pretend-decision-sharing than pretend-ownership, i.e. businesses that are egalitarian on paper but not in reality. Europe is full of "service co-ops" where people are members in theory but simple employees in practice, orgs built just to game taxes or minimum-wage regulations. That's the worst.
These days no one really envisions working for the same company for 30 years. It will be tough for leaders within Zappos to show on paper their leadership progression.
I think the idea of holacracy has legs, but I haven't seen any other implementation that would lend credence so far.
Holacracy to me sounds like Agile or DevOps. Maybe one or two places are doing Agile/DevOps correctly and everybody else is just doing some kind of bastardize form of Waterfall, or as they call it at my place, "AgileFall".
This may turn out selecting for the right people. I’ve learned as a manager in a growing company that a lot of people really like traditional hierarchy, and for understandable reasons. I’ve learned to respect and work with that. It’s not my way, per se, but I get it.
But from my reading it sounds like the truth is at least partly: by fudging it.
Spreading it out more equitably among all employees sounds good until you realize that it makes decision making even more opaque: when everyone is responsible, nobody is.
Indeed; and in fact it is now obligatory to mention the excellent Essay of ancient wisdom on the subject: The Tyranny of Structurelessness http://www.jofreeman.com/joreen/tyranny.htm
Having a flat customer service organization (and valuing making customers happy) sounds like a great way to filter for reps who can think independently and find ways to make customers happy (and themselves more efficient).
I'm actually surprised they don't see more people gaming the system, applying to get the bonus without any intention of working at Zappos.
That being said, the new setup is great for people who are used to that level of independence. Were you a freelancer? In an early stage startup? Came from schooling where you were in charge of your education? If that's you, the new Zappos is a natural fit. It also puts you in a tiny (lucky) minority of the potential candidates who are.
This is like posting an article asking "why are so few people smoking in restaurants?" right after a no smoking rule was introduced in restaurants.
Unfortunately, I do think they will run into some adverse selection issues. The best employees are also the ones who can easily find a new replacement job and collect a free 3-month exit bonus. Not to mention that many top performers probably rose into management and are facing the possibility of a pay cut.
Holacracy does not equal an absence of hierarchy or "flat" workplace.
this latest round of news stories all stem from the offer that was given in may being extended to a select group who were deemed critical to finishing the "supercloud" migration. that extension ended in january, so another "exodus" occurred as a new round of people took the offer. since the project was a tech project, this new round was almost 100% from tech.
the ones who accepted this latest round were heavily tilted toward managers. mostly product or project managers with a healthy smattering of engineering managers. out of all the engineering managers (~15), there are only 2-3 of us left. we are the ones who took on active roles in writing software and making an impact by doing rather than commanding (https://medium.com/zappos-engineering/elastic-beanstalk-vs-a...).
fwiw, from my perspective, the story is far from over. i feel like it is only now that we can really progress the hard work of self-organization. the extended offer was holding us back. it's tough to move forward with distributed authority when you know that a good portion of your peers are not interested in that distributed authority.
i guess the message i'm trying to get across now is that for those of us here and moving forward, it's far from a "finished" project. it's certainly not a "failed" experiment yet. it feels like it's just starting.
and we are starting to see some of the benefits of people self-organizing. one example is that for the last couple of years we've tried to contribute to open source. i'm not entirely sure why we weren't able to do so except to point at bureaucracy and traditional management putting up roadblocks. if your manager doesn't see how you working on open source is going to help them out, then you have much less incentive to contribute to open source.
in a self-organized system, with the directive being "figure out what the best use of your time is for the purpose of the company," when an individual senses that contributing to open source benefits the company, then they can do it. since the tearing down of the traditional management hierarchy, we've had an open source project released and are deep into releasing a second.
it's hard to say this never would have been possible under the traditional hierarchy, but it is accurate to say that it didn't happen after two years of trying and then happened within months with self-organization.
A high turnaround was to be expected. The interesting part would be if things stabilize again.
I think we have enough proof that the flat management structure doesn't work as well as a traditional hierarchy for a company of Zappos' size and in its industry. But I guess some people won't believe it until the company actually goes down the toilet or backtracks.
> But I guess some people won't believe it until the company actually goes down the toilet or backtracks.
You fall under that category. The criteria that Zappos website has to actually go down in order for it to have been a bad decision is a pretty bad way to make a judgment call. If you ran a company that way you'd never fix a bad decision until it was too late.
you're making many mistakes in your opinion:
1. correlation and causation. this latest round of people leaving is only because there was an extension to the offer made in may for a group of people. far less left this last time than anticipated.
2. you aren't running the test long enough. because of this second offer, the hard work of self-organizing was largely being delayed. it actually feels like we are just beginning (at least in tech).
3. you are measuring the wrong thing. in no way does an attrition rate reflect the success of self-organization. any organization can be made better byt he right people leaving.
the metrics i'm interested in are: 1) employee happiness, 2) customer delight, and 3) innovation rate. feel free to email me in a year to get an idea of what those are like then. (notice, a profitability metric is just an assumed by-product of those measurements, although it's not what we focus on).
Managing people is a complex problem that requires a specific set of skills which take time and practice to get right. I've known (and reported to) incredible managers who not only got us to achieve more than our team would have otherwise, but made us enjoy our jobs more in the process. I've also had managers who've lacked in that skillset, and had the exact opposite effect. And it was never because they were assholes or sociopaths- they just didn't have the skills they needed for the job.
I may take on the management role someday, not because of some dream of lording over people and giving orders (that part actually turns me off) but because it's a new domain of complex problems.
But the roles does come with some power. And after a while, it is human nature to get used to that privilege that it starts becoming invisible.
Getting used to being a manager (or the 'privilege' as you call it) is not a bad thing and in no way does your interpretation of human nature make any difference.
I wrote that the powers that come from being a manager is a privilege. Are you arguing that it is not? Or are you arguing that managers don't have such powers?
The "no true manager" argument is always brought out when I say something critical of the management profession.
>Power isn't the benefit of the role unless you're a sociopath or an asshole.
I agree. Guess who inevitably fills management jobs at medium to large companies.
For instance, I just left a well paying job at a very large company because my manager quit: he though no initiative could never be successful in the internal political climate, so he left despite being offered a big promotion. Since he left, the company had 6 high profile resignations, as we all knew the alternatives the head of IT was considering were just bad managers. Now, with tentpole technical people leaving, a lot of other qualified people are looking for new jobs too: It was quality leadership that was bringing the team together.
Really, your original comment on this topic would be 100% fair if you were describing what this company is going to end up with. I have worked for horrible, horrible managers over the years. It's just that once you have seen good management, and understand the wonders of productivity one of those can achieve, it's just very hard to agree with a view quite as dire as yours.
There's all this companies that are working on helping recruit good developers: In my experience, good developers are not really THAT hard to find: But a company that could help hire good managers, and properly identify the people that have management qualities in our office, as opposed to just those that look and sound managerial? Now that's a company that would be worth hiring.
That's a fair point- I won't say there aren't any bad managers- there are sociopaths attracted to the field. But I don't think you've met any of the good ones either. I'm sorry to hear that.
> Guess who inevitably fills management jobs at medium to large companies.
The sociopath-style manager will only have short-term gains compared to the managers who've acquired the skill set for the domain. One's people will be unhappy and quit; the other's will stay and encourage friends to join. After a few years of this, when upper management want to promote to a manager-of-managers, the long-term results will speak for themselves.
That's how I lost my best manager- he succeeded so widely that he was promoted up another level (where he continued to do very well).
My advice to you is to consider how you would behave as a manager. How would you run a team of ICs such that they'd be happy, productive, and meeting all goals? It's not about whether you'd want to or not, it's how you'd approach the problem.
No they won't because corporate management creates perverse policies of continuous turnover and stack ranking for this exact reason. Losing all your people many times over in a short time is just portrayed as a sign of a good manager culling the herd.
After all, if I was going to do what I think is profitable, I don't think accepting a pay check from the company is worth giving up the entire upside of my own ideas - I'd quit and do my own thing. If I knew what customers wanted, and the company didn't, why stay?
You could have a situation where the sales people talk directly to the developers. If sales team were the managers, then they tend to over-promise. If engineering was in charge, then what they build tend to be out-of-touch with their market.
Taking out the manager is to bet that sales team and engineering team can work with each other directly with equal clout, and be able to maintain that equilibrium.
But for my own situation, if the founders decided to quit and promote me (and everyone else) to lead the company as a holacracy, providing only salary and no equity, I'd quit. I could work on my own ideas by myself with all of the equity, why do it in a company where I have none?
If you're hiring makers and doers who don't value good management, you will not be able to function.