People don’t leave companies. They leave managers.
I’d put that in quotes, but I can’t find an attribution. It’s become such common knowledge, and it’s been written about so often, that I think everyone who is interested in employee morale and the benefits of keeping your good employees happy already knows this.
This statement really rings true for me. But I don’t think it rings true the way that it’s often meant. I think that when most people hear that, they picture an overbearing, boorish, emotional, or abusive boss, pointy-haired or not. They maybe think of the condescending boss that uses this type of conversational antipattern.
Really, though, I’m thinking more about management structures today. I think people are just as likely to leave management philosophies as they are to leave individual managers. Sure, having a terrible supervisor has the potential to make every moment of your working life terrible, but the way organizations shift and rotate these days, having to report to a duff manager for a while isn’t really that big a deal. Kinda like the old saw “If you don’t like the weather in <remarkable number of regions/cities>, wait 5 minutes and it’ll change.”
Bad management philosophy can break a company, one spirit at a time.
And yes, in the past couple decades, we have seen the old command and control style of management from the industrialization era falter. We have seen the rise of agile project management and lean software development. We have watched with curiosity at the mounting evidence that waterfall-style management of software projects and programs does not take into account the pragmatic reality under which software is routinely delivered.
There are still lots of companies out there whose leadership hasn't received the memo on this, yet; mainly those whom were educated under the philosophies and by the leaders of the industrial era. There are a lot of mid to large companies whose leadership hasn’t picked up a management book in twenty years, to hear people describe their environments.
I want to discuss a few things that I think would be massively beneficial to any company currently stuck in the old, rigid, empire-building, command and control management style.
Resource rotation
One of the big antipatterns I’ve seen in organization is resource-hugging. That is, there are five developers on Team A and three on Team B. Team B gets an organizational-critical project, which would really benefit from having another developer. Let's say that no one disputes that Team B’s project is more important that Team A’s. And let’s say that no one disputes that it would be a better use of company resources, even Team A. What are the odds that Team A will lend Team B a resource? Factoring in not only willingness, but also organizational red tape? I’m guessing it’s close to zero. Managers might like to think of themselves as working for the good of the organization, but silos and affinity can run deep.
Old-school organizations have a really difficult time with letting developers switch products and project managers, despite the positive benefits that come from such rotation. Some inhibitors to this type of rotation include siloed teams not having similar enough processes, teams having completely different coding styles, needing visibility for annual performance reviews, and so on.
More modern management philosophy suggests that rotation through projects, products, and managers engenders similarities of codebases, commonality of processes, and dissemination of knowledge. It reduces the number of junk drawers in the organization, and limits loss of knowledge due to lossy transition plans.
Also, while this may be a post for a different day, I think there’s a huge case to be made for rotating your most senior IT leaders. Especially in small shops where there are folks that have been there a long time, and have very deep relationships with the business. Sure those relationships are optimized in a way, but they’re not optimized for global company benefit, and let’s face it, the whole company is what’s got to survive and thrive. It does no individual silo any good to succeed if performance problems in another silo drag down the organization.
Flat hierarchies
Another big no-no in modern organizations is the steeply peaked org chart. While there’s some evidence that extremely wide spans of control (1 manager to 20-30 people) can get out of control, there’s also evidence suggesting that a company loses a lot of communicative efficiency when there are a lot of levels between the lowest-level leaf-node worker and the CEO.
Having steeply peaked hierarchies leads to another big issue, and that’s one of throughput. If an organization has a high management to staff ratio, it’s going to have a bottleneck. Especially if there are non-people-managing project managers in the mix. Each manager in the organization is given at least one project that is their top priority. They may need a couple FTEs on the project and other ancillary folks. With a steeply peaked hierarchy, there are not enough leaf-node workers to get everyone’s project done. The managers then have the ability to say, “Well, I asked for some work to get done, but another manager’s project came first.” In the worst case of this, the work is not prioritized actively and the whims of what the developers want to work on, or the manager that barks the loudest drives what the organization accomplishes.
It seems to me that huge efficiency that would accrue to any organization that would recognize the preceding and deal with their jelly layer (mainly by having a lower jelly to producer ratio). Most organizations need more driven, informed individuals doing work, and fewer people arguing about why the work isn't getting done. Scott Berkun suggests a great way to deal with the "too many chiefs" syndrome: innovation by firing people. Seriously, if your organization is so top heavy that you've got more projects than people, start there.
I also am really fascinated with the holocratic experiment that Zappos is conducting right now. I’m far from declaring that hierarchy should completely dodo, but I think current data suggests that flatter is better.
On a side note, steeply peaked organizations with deep siloed hierarchies also really suffer when coupled with ineffective delegation. Every level of the organization should maintain at least some spending limit that, below which, they have absolute authority to make a call. A $30 software tool to improve developer productivity should not have to go all the way to a CEO or even the head of technology for a decision. Before you've even described the problem, you've already spent more of the company's money in wages discussing it. When you don't have large enough spending approval limits low enough on the org chart, you can run into the longest delays for the simplest things, because validation has to make it through so many more layers of jelly.
A parting thought: one of the reasons that I've often heard for narrow spans of control is that the annual review becomes too hard to manage for one person. This is a really bassackwards reason for having a peaked hierarchy, as it gutpunches you twice! Peaked structures have awful consequences, especially coupled with inefficient delegation, and every time an organization narrows spans of control solely to participate in one of the worst business practices, an angel gets cancer.
Differentiation between managers and team members
A manager cannot be held to perform the same tasks as their employees. While managers may have come from the ranks of the individuals that report to them, and are invaluable in terms of providing overall direction, having a manager that currently performs the same tasks as their team members is a roadblock to innovation and prevents productive dissent.
Here's how. Let's say two people are doing the same job, only one is manager of the group, and one is a team member. If the team member has a difference of opinion with the manager, that team member has a limited set of choices: disagree with the manager publicly and risk getting reviewed poorly or being disciplined (in public or private), or play along and let a potentially bad decision for the organization play out for purely political reasons. Bad things happen when team members are set up by organizational structure to compete. You either end up with no innovation or people refuse to continue playing along and depart.
If two people are fulfilling the same role in the organization, they are team members. If there's not enough management work for full time management duties within the group, consider making a team member a team lead for administrative purposes, but don't change that reporting structure. That's poison in the morale well. Pursuant to the previous point about flatter = better, be very sparse about where you decide to install organizational jelly layers.
Make sure the org change will be welcome
If you’re going to ask people to report to someone new, you should check with them to find out if there are any objections. There’s nothing more impactful to productivity than an employee not respecting the person they suddenly need to report to. This doesn't have to be a big formal ordeal; even something like, "I'm thinking about restructuring the organization so that you'd be reporting to <whomever>. How do you feel about that?"
Springing a restructuring on people is disrespectful and does not engender trust. It makes it look as if the organization has something to hide, as if discussing it with employees beforehand would somehow queer the deal. If it's such a great idea, it will be embraced. If there's going to be a problem, you want to know that up front as an organization. Also, if you're going to lose people as a result (because people do leave managers), make sure they're people you want to lose.
This holds true for new hires, by the way. I can't articulate how important it is for an incoming manager to meet the team they'll be working with, even if as a final "sniff test". Too many organizations only have interviewees interview up. In many modern business organizations, if they do reviews, HR requires 360 reviews. It only makes sense to offer the team that you're hiring a manager for to meet the interviewee in kind of a 360 interview.
R-E-S-P-E-C-T
The reporting line in the organization should form an unbroken chain of respect. That is, anyone asked to support a manager must respect that manager. If they don't, the entire organization suffers. You can ask someone to report to someone else, but if they don't respect that person or their abilities, there's going to be trouble. Lack of respect for someone the organization trusts to manage leads to lack of respect for the judgment of the organization.
The worst place that this lack of respect can show up is in promotions. Everywhere I've ever worked, I've seen people promoted - not beyond their abilities, because sometimes you have to trust people to stretch and grow - but beyond their personal charisma and respect level. Promoting someone who does not engender respect or trust forms a weak link in the leadership level.
The worst place that this lack of respect can show up is in promotions. Everywhere I've ever worked, I've seen people promoted - not beyond their abilities, because sometimes you have to trust people to stretch and grow - but beyond their personal charisma and respect level. Promoting someone who does not engender respect or trust forms a weak link in the leadership level.
And respect is one of those things like trust that cannot be demanded - only earned.
Conclusions
Ultimately, a strong and confident organization puts leaders in place that engender respect, not through level or title, but through good decisions and good people skills. Given that you can have leadership at all levels, consider policies that empower them to use their judgment for the good of the company. And remember, because management != leadership, you don't have to have a million managers in your company to get a lot of great work done.
Conclusions
Ultimately, a strong and confident organization puts leaders in place that engender respect, not through level or title, but through good decisions and good people skills. Given that you can have leadership at all levels, consider policies that empower them to use their judgment for the good of the company. And remember, because management != leadership, you don't have to have a million managers in your company to get a lot of great work done.