Thoughts on Co-Leadership: What Does Success Look Like?

This is the third of four posts. Each tackles a piece of the elephant that is the recent Fractured Atlas move to a four-person “Chiefs Executive.”

The opinions expressed in these posts are entirely my own and do not necessarily reflect the views of any other Board member, the Fractured Atlas organization, or its staff.

If you missed the other posts, check them out here: part one; part two.

What to Measure and When to Measure It

What will success in this model look like? We’ve certainly discussed that question as a Board, often and intensively. What’s lacking, in my view, is a clear path to measurement. I can’t solve that problem in this post, but I want to highlight some of the challenges I think we’ll face and suggest ways we might overcome them.

As a Board, we’ve supported co-leadership based on two decision factors: (1) our commitment to anti-racism, anti-oppression (ARAO) leadership; plus (2) an instrumental hypothesis that a single CEO was a performance bottleneck for the organization, and that a four-person CEO would improve greatly the organization’s “executive bandwidth” (understood as an amalgam of intake/throughput/result) for dealing with complex resource flows within the organization and without.

The first factor, ARAO impact, is a long-term project whose results likely won’t be visible quickly. We know how to measure it within our own programs and services but measuring the externalities is going to be very challenging.

(For reasons discussed in an earlier post, I’ll leave the ARAO discussion to my colleagues and focus on the second, instrumental motive here. Please do not conclude from this that I consider ARAO a ‘secondary’ benefit of this work: I’m taking a lesson from our co-leaders and sticking to my area of relative competence.)

The second, “instrumental” factor is both vital to our immediate and long-term sustainability as an organization and important to our ability to influence others to seek the same ARAO benefits the same way. It commits us to the hypothesis that success looks like (much) greater executive bandwidth. Our theory of change is that improved executive bandwidth will improve meaningfully one or more of our four key strategic, organizational outcomes: mission-performance, agility and resilience, self-determination, and sustainability—net of the costs of change, of course.

Building a Better Bottleneck

As we think through this work, I keep coming back to one underlying insight stemming from the idea that organizations are systems of complex contracts (see post #1 for a discussion of this concept). Every real-world system has a “bottleneck:” a constraint that limits the impact the entire system can deliver.

This idea of bottlenecks also lurks inside the idea of relationships as complex contracts. Different types of contracts accommodate different quantities and qualities of resource flows. The wrong contract in the wrong place at the wrong time can produce a substantial bottleneck in something—information, money, effort, other resources, or decisions—that the organization needs to thrive or even to survive.

When one eliminates a bottleneck, systems don’t become bottleneck-free; instead, a new bottleneck takes its place. The restrictiveness (“size”) of the new bottleneck determines how much value one gains from eliminating the old one: if the new bottleneck lets a lot more through, the gain is large; if it’s the same size as the previous one, the gain is nil.

Bottlenecks cost resources to remove; some cost a lot. Whether we gain enough to make the cost worthwhile is determined by the restrictiveness of the new bottleneck.

In real life, bottlenecks can occur in any sequence and be very far from one another both physically and logically. Removing the very first bottleneck can solve almost all the problem, or one may need to remove hundreds before any significant gains appear. To navigate this challenge successfully, the key insight is that one must evaluate the system as a system, then re-evaluate it, after each new bottleneck is discovered and removed.

The process is iterative and subtle: it’s usually impossible (in any moderately complex, real-world system) to know what the next bottleneck will be until you remove the present one and then re-measure the entire system to see what’s now constraining it. For more on these concepts, see the work of Eliyahu Goldratt; start with The Goal.

Finding the Next Bottleneck

If we consider the move to our new, four-person model as the removal of the first bottleneck (the single CEO), we can improve its chances of success markedly by finding the next, new bottleneck on executive capability and, if feasible, eliminating it.

The more bottlenecks we can discover and eliminate, the greater the chance that the gains from our experiment will be large enough to make the new model viable. If we can’t eliminate the next, new bottleneck, and if we learn it puts a low ceiling on the gains we can expect, we will know that the promise of our model is largely illusory.

Let’s put this in more-concrete terms. By moving to four co-leaders, we’ve roughly quadrupled (minus the collaboration costs, which are not trivial) the intellectual processing capability of our executive function. Does that mean the co-leaders will be roughly four times more effective than a single executive? Unfortunately, not necessarily.

After subtracting the collaboration costs, any gain will be dependent on how much additional result they deliver in fact, not on how much they can theoretically deliver. The increase in result, in turn, is determined by the location and nature of whatever is the new bottleneck on executive capability in our new system. That could be anywhere and anything, from some technical limitation that’s easy to fix, to the human need for sleep, which is not.

Logically, the new bottleneck can be found in one of four places:

• Collaboration costs (a bottleneck internal to the relationships among the co-leaders). This is a very likely locale for the new bottleneck, as it takes a lot of effort to keep four people fully informed about each other’s activities and even more effort to give all of them what they need to allocate and make the right decisions.
  
  If the next bottleneck is found here, and if it’s severe, steps will need to be taken to relieve it. The co-leaders continue to work at this, and the Board continues to monitor that work.
  
  
• Organization-internal. It’s possible, for example, that the elevation of four co-leaders will prompt changes that drive so much more organization-internal information to the co-leaders that any gains are lost to the “new noise.” Or, the changes could be independent, like a coincidentally timed move to a purely work-from-home organization, but still have the same effect.
  
  If the new bottleneck is found inside the organization then, by definition, the co-leaders have both resources and formal authority to address it, reducing but not eliminating concerns.
  
  
• Environmental. Given the level of connectedness enjoyed by FA and the tremendous intellectual bandwidth of today’s professional networks thanks to social media and the World Wide Web, it’s difficult to imagine that the new bottleneck lies somewhere fully outside the organization.
  
  We’re likely to reach the individual information processing limits of our four co-leaders long before we exhaust the fire-hose of information flowing into our organization from the world outside. If there are bottlenecks, however, they lie entirely outside the organization, so we might not have any ability to address them.
  
  
• The Board. After collaboration costs, this seems to me to be the second most-likely place to look for the new bottleneck—and if we do find and remove a bottleneck in collaboration costs, it seems highly likely to me that the next bottleneck will be found somewhere in our Board.
  
  Our Board operates on current best-practices, meaning that it’s optimized for a single CEO. Theoretically, it would be remarkable if Board institutions and practices optimized to serve one individual worked just as well to serve four. In practice, we’ve already accumulated ample evidence that they don’t.
  
  Board communications have not ramped up to match the ramped-up communications among our leadership team. Episodic Board engagement—a best-practice for many nonprofits—introduces severe lags in information flows and decision timing. These are serious bottleneck candidates.

In talking about bottlenecks, I’m using an engineering metaphor more-suited to water or electric systems than to complex human organizations. In an organization, finding and measuring the bottleneck isn’t simple. What does “executive effectiveness” look like, and how do we know when it’s being blocked? These are hard questions in any organization. In for-profits, the concept of a “bottom-line” helps tremendously. When an organization’s primary currency is “mission attainment,” the task gets even harder.

Precisely Imprecise

In practice, we can’t measure with any precision. Instead, we do something closer to eyeballing bottlenecks and addressing the one(s) that seem to interfere the most. We’ll most likely find several and try to address them all, never knowing which one was “the” next bottleneck. That’s what the co-leaders are doing with collaboration costs and, I would argue, that’s what we need to start doing as a Board with our own structures and activities. Precision would be lovely but, in its absence, we need to press on.

With this understanding, it’s time to turn to the question of what we as a Board should be doing, going forward, to support our experiment and give it the best possible chances of success. Where should we be bottleneck-hunting, and what should we be prepared to do when we find some?

The topic is large, diverse, and important. Our Board Chair, Russell Willis Taylor, joins me in post #4, coming soon, to give it full measure.

For more thoughts from the Fractured Atlas Board of Directors (and to read the other pieces in this series), click here.