Dustin Moskovitz, the co-founder and CEO of Asana, wrote a cautionary piece on Medium (see Goldilocks Management) about the growing trend of zero management, as typified by Valve. He starts by posting the intro to Valve’s Handbook For New Employees, Welcome to Flatland, which says ‘nobody “reports to” anybody else’, and that the company has dropped top-down, military-style command-and-control organizational decision-making.
Moskovitz then hops to the end of the Handbook, where the down side of this zero management approach is laid out [note that the marginalia is in the original copy at Valve's website]:
This is where Moskovitz draws the line, and where he thinks the tradeoffs are too high:
In exchange for freedom, they gave up *personal growth.* I can fully believe that there are people who can survive in such an organization, but I am deeply skeptical that they can thrive and reach their full potential. They are throwing the baby out with the bathwater. They tied one hand behind their back. The juice is not worth the squeeze. Pick your favorite metaphor; the point is this philosophy lacks balance. More importantly, I totally reject the idea that freedom and mentorship have to be mutually exclusive. Many people believe Valve succeeds thanks to this culture, but I suspect they actually succeed in spite of it and would be a better company if they backed off from such an extreme position.
I agree with Moskovitz in part. I don’t think that people should have zero mentorship. However, my sense is that even in very conventional organizations, mentoring is not something that is necessarily undertaken by a first line manager. It is a unique skill, and something that managers may not have the skills or inclination to perform.
In an organization tending toward what I am now calling leanership, where a high degree of autonomy and self-direction is the norm, individuals may not have a direct reporting relationship of the conventional sort. The various elements of a ‘manager’ may have become disentangled, so that some are handled by diffused networks — peer-based performance reviews, for example — while others could be performed by non-managerial specialists.
Moskovitz relates the take of Josh Smith, a new grad who joined the company and did well:
Our employees also value that our approach emphasizes personal growth, especially through mentorship. One example of someone who benefited from this is Josh Smith, who joined us as a developer straight out of college, after first doing an internship. He already was great at some things related to his role, but will admit he was lacking in many others. Josh said that his manager, Jack, pushed him to test his boundaries in these areas specifically, “at a speed where I was constantly challenged, but not overwhelmed.” This ensured he grew quickly. After just a little less than a year as a full time developer, Josh has gained enough institutional knowledge and experience to become a program lead and begin playing the mentor role himself.
Consider Bette, a new designer hired into a leanership product organization. Rather than a direct report, she has a number of relationships with distributed areas of involvement in her work. One is an old friend, Carl, who had recruited her to the company, although working in customer support. Bette has asked Carl to mentor her in the ways of the business, and Carl has agreed, and that relationship has been made public. Carl has made it a point to discuss the company’s approach to human factors, and to make several key introductions even before she was hired.
But Bette is a high performer, and not a new graduate. She has made a deep commitment to her own work, and has worked long and hard at mastering her skills as an industrial designer. That doesn’t mean she is past all improvement, and has no need of mentoring. But she is engaged in a long-term investment of time and energy in refining her craft, as the central aspect of her work identity.
She also has a primary working relationship with a group of designers working on a new wearable device, the team where she was hired. In this organization, all the members of the team have to agree on a new hire, and she worked for a few weeks on a part-time paid project so that they could all get to know each other (see DuckDuckGo’s Gabriel Weinberg talks about ‘Inbound Hiring’). The team runs lean, and various specialities are merged in the group — so Bette has the longest experience in industrial design, while others are specialists in apps and operating systems. It is a cooperative group in which decision making is both shared and owned. For example, questions about the hardware design are likely to be resolved by Bette — since that is her area of expertise — and iOS design questions would most likely be answered by someone else.
But what if Bette and the others disagree on some key issue, and it cannot be resolved easily? In a traditional organization there would always be a manager of the team who would step in and make the call. In fact, one of the most pernicious aspects of traditional slow-and-tight companies is that managers often make as many decisions as they can, stifling initiative and bottlenecking innovation.
But Moskovitz explains Asana’s approach, which is less conventional, and where the resolution of a lack of consensus in a design issue propagates outward like exception handling in a programming language. In Asana, the company defines so-called ‘Areas of Responsibility’:
An alternative approach to creating structure and order that we employ at Asana is distributed responsibility, exemplified by our AoR (Area of Responsibility) program. Rather than have all decisions flow through the management hierarchy, we have the explicit intention of distributing them as evenly as possible across all employees. Unless we forgot about the existence of an AoR (which does happen) or otherwise make a mistake, the relevant domain owner has the final decision making authority, not their manager. Additionally, we try to promote an understanding of management as an AoR on par with other AoR’s, making the organization feel flatter.
Management still plays a very important role, however, as a backstop for all decisions. If an AoR does not exist or the most relevant decision maker is ambiguous, then decisions do flow through the management hierarchy.
In a leanership organization, there might be zero management hierarchy, and only areas of responsibility. Asana is using management as a quick way to resolve such calls, while in a learnership organization the foundational constitution would lay out how responsibilities are defined, and then the organization might be involved in an on-going process of assigning responsibilities.
When Bette was hired, she was specifically assigned responsibility for hardware design on the new wearable. Now, given the design issue with the iOS designer, the issue would be propagated out to the next layer of responsibility, which is falls into the hands of a senior designer, Bryan, one of the company’s founders, who has been keeping a close eye on the project. He agrees with Bette’s approach, although suggesting a different angle for the iOS issue than had been considered, based on work Bryan had done on an earlier iOS project. Bette and the iOS designer work out the details, and the issue is resolved.
Some might say that this layering of responsibilities in a leanership system is just a hierarchy by another name, but it is not. Bette doesn’t report to Bryan: he’s not the one responsible for her salary review, or tasking her with new work, or any other aspect of management. He’s not her mentor.
The set-up works more like the separation of powers in the US system of government. At different times people play roles defined within legislative, executive or judicial parts of the constitution. Bryan’s involvement in the design disagreement has him acting as a judge resolving a dispute, and not like a superior in the executive branch of the company. And if the company starts to see a greater number of these sorts of disagreements arising, the company might have to rethink the layers of responsibility, which would be like the legislature passing new laws, approving new owners of responsibilities, or modifying the constitution itself.
But in a leanership company, everyone is part of the executive branch, almost all of the time, and only assume judicial or legislative roles occasionally, as needed, on demand. That’s why the company runs lean: because much of the work that traditionally falls to full-time managers is instead handled by exception by people who step into the role of an arbiter, and then go back to their day job.
This capacity to take on and then relinquish authority is called emergent leadership by Lazlo Bock at Google (see Lazlo Bock talks about hiring at Google, and why the GPA is irrelevant), and is the centerpoint of leanership organizations, where largely autonomous workers share authority, and conflicts are handled by exception in expanding layers of responsibility.