At the Uncertainty Project, we explore models and techniques for managing uncertainty, decision making, and strategy. Every week we package up our learnings and share them with the 2,000+ leaders like you that read this newsletter!

In case you missed it, last week we talked about rationality and biases in complexity and uncertainty in a guest post by Marco Valente.

Designing for aligned autonomy

This idea of the science of design is defined, outlined, and explored in Herbert Simon’s landmark book, “The Sciences of the Artificial”. Loosely speaking, it’s his term for the engineering disciplines - those efforts that create “artifacts”, in pursuit of goals. 

Here is Simon’s concept of the artificial world, bringing the focus to where we “devise artifacts to obtain goals”:

“An artifact can be thought of as a meeting point - an ‘interface’ in today’s terms - between an ‘inner’ environment, the substance and organization of the artifact itself, and an ‘outer’ environment, the surroundings in which it operates.”

“The artificial world is centered precisely on this interface between the inner and outer environments; it is concerned with obtaining goals by adapting the former to the latter.”

In this post, I’d like to go deep on parts of his chapter on “Economic Rationality”. Call it a book report, if you’d like. Or a long game of “Simon Says”... <groan> 

While the book covers a LOT of ground, I particularly liked the sections focusing on organizational design. He called it out as just one example of an artificial system that we design and create to pursue our goals. I’ll focus on this area, since Simon offers great insights on questions we have explored here at The Uncertainty Project, namely:

• Where and how to make decisions in an organization, when faced with significant uncertainty

• How to align individual, distinct efforts and choices, when decentralized

Simon sets the stage by drawing a contrast between decentralized organizational designs that emphasize agency and autonomy at the edges (i.e. market-based), and centralized organizational designs that emphasize planning and hierarchies (i.e. organization-based).

While market-based structures are well known for their efficiency, as they coordinate decisions and behaviors of independent actors, organizations are usually seen as inefficient bureaucracies. In contrast to markets, organization-based structures rely on central planning, often based on statistics, to coordinate efforts via hierarchical structures with formal authority that establishes communications patterns up, down, and across the leadership ranks.

These two distinct styles of coordination, market-based and organization-based, are mixed together in our modern business environments, as we look to leverage the advantages of both.

Let’s imagine designs at the extremes:

• In a purely market-based company, you would arrive at work each day as an entrepreneur, or gig worker, establishing contracts for your services on a dynamic basis. [Sounds exhausting!]

• In a purely organization-based company, you are a “resource” directed to perform specific work by people up the hierarchy, without much creative thought and contribution to the plans. [No thanks!]

Simon explains why organization-based systems can be more effective in some cases:

The stability we gain (on both sides) from the employment relationship, is a better “artifact” for us to use to pursue our goals.

But we trade this stability for a new challenge, related to incentives. Markets gain their efficiencies via their ruthless (self-centered) focus on pursuing goals in our personal interest. When this is lost, something else must point the way.

These are the same challenges we see and hear about in our modern workplace.

• How can we create motivated, engaged employees?

• How can I create aligned autonomy with shared visions and incentive structures?

• How can I build innovation- and entrepreneurial- mindsets into our culture?

So, given these challenges, what’s in our toolbox, for organizational design? 

The tools we have available can be categorized on this spectrum from market-based (autonomous agents) to organization-based (centralized and hierarchical). Your goal is to build an organizational design that positions your part of the company in the appropriate spot on the spectrum, given your context, or surroundings.

Simon points out that organization-based structures get a bad rap.

Specifically, he defends ways that organization-based structures can be more efficient than just relying on market forces:

• When dealing with externalities

• When dealing with uncertainty

• When leveraging the power of identification to drive alignment

Within large corporations, we often lack “true” market pricing for the exchange of value between one group (or team) and another. While it might work in theory, it is elusive in practice.

This is the realm of annual budgetary planning, where funding is doled out based on negotiated understandings of these internal markets. This is how the internal externality (i.e. no real market pricing) is addressed.

Another area where organization-based structures can have a leg up is in their ability to stay aligned in the presence of uncertainty.

This ability to socialize and share a set of assumptions and predictions stands in stark contrast to a market-based approach, that rewards the independent agents making good assumptions in a “survival-of-the-fittest” style. 

But who wins if one part of the company makes good assumptions while another part makes bad assumptions (about the same question)? Not the company overall, that’s for sure. We form large organizations to help individual actors achieve more than they could alone. But aligned action requires aligned assumptions, and hierarchy can help with that kind of coordination challenge.

What kind of uncertainty (or what magnitude) should drive us to pull “hierarchy” out of our toolbox?

To answer that, we need to explore the sources of the uncertainty.

This feels like common sense: when business units or product lines are serving different markets or customers, then we can decentralize the decision making (and assumption making) to those distinct parts. The need for aligned action is less than cases where the uncertainty affects all the parts in similar ways, like a global economic downturn, or a pandemic, or in cases where multiple products serve the same market or customers.

When decision making is decentralized, we gain efficiencies from the independence of the actors, acting as a market-based system. But what other tools can we pull from the toolbox when the sources of uncertainty drive us to use some “hierarchy” in our structures? For example, what incentives support alignment across the organization, when personal interest can no longer be the primary motivator?

For this, Simon points to our (very human) need to belong.

When working in high-performing teams, we align, as members of the team, around the common goals. This is the super-power of a team, and it relies on this voluntary choice to identify with the team. This is why we wear the swag, enjoy eating our two-pizza lunches, and (hopefully) enjoy seeing people at a standup meeting.

The sweet spot happens when individuals are able to shape a shared vision, working with the group. As Peter Senge highlighted, when an individual embraces a shared vision, it still takes shape as a personal vision; it just happens to overlap, or be common with, others’ visions. 

But our personal visions are blurry, built with incomplete information. We have to create a simple model of our complex world, and this feels like an easier task when we build that model with others, pooling our efforts, and shaping it collectively, based on collective interests.

Note this self-serving benefit of identifying with a larger group (i.e. team or organization). We learn faster. We are more likely to succeed, due to this learning, than if we were independent. This makes up for the loss of individual autonomy.  

Recently, this kind of loyalty has been under some strain, as we adapt to our new remote and hybrid work environments. Human nature says that proximity influences the strength of our loyalty.

Have you noticed any shift in your loyalty, when you shift your location? I feel like I have. When I’m home, my loyalty to my family is a constant presence. When I’m at the office, I more easily recognize my loyalty to the firm, and more easily build this additional identity alongside my colleagues. 

But this identification is a choice, made by individuals. Simon uses an interesting term to describe this choice to accept information from others (to learn faster) and to be swayed by social influence. He calls it being “docile”:

So whether you like the term ”docile” or not (to me it sounds passive, acting like lambs instead of wolves…), this is a key insight on the tradeoff we are making - constantly - in our sense-making. 

Does this resonate? Do you feel like you (sometimes) willingly accept information from others, because it’s easier than researching it yourself? We all do it, all the time. It’s the expression of trust. It’s also a survival skill.

But making this choice can also be a source of conflict - sometimes healthy, sometimes not. When you stubbornly resist agreement in a social setting, you are deciding that your “fitness” is better off, if you plot your own course. You’re gonna go touch that hot stove, dammit. This is the messy dynamic of group decision making.

So “docility” can be a good thing. But is there a point where it could go too far? Is there a point where we are accepting information in situations where we should be challenging it? 

Yes, and this choice is framed by Simon as opting-in to pay a “tax” (as a docile individual who is receiving this benefit of more information), which is imposed by the organization.

So it is this benefit of alignment, without a significant loss in incentive, that Simon feels is the magic of organizational-based structures.

Hierarchies (short ones) with (some) central planning are not inherently a bad thing. Pulling these organization-based tools from the toolkit can be warranted when:

1. You need some coordination across independent activities, to keep individuals from trying “to outguess one another”

2. Individual efforts can’t easily be tied to individual rewards (the “public goods problem”)

3. Decisions require information that should be coordinated across individuals, teams, or groups

Simon makes a case that in an organization-based structure, instilling organizational loyalty is an inexpensive way to incentivize the organization’s goals over the incentives of personal interest. And as the late Charlie Munger told us, we need to be paying attention to these rewards and incentives, if we want to understand and influence the world around us.

As you explore ways to decentralize decision making and empower teams, consider these insights from Herbert Simon about centralizing common assumptions and using identification to incentivize around shared visions.

Join us this week for the Decision Architecture discussion series!

We’ve set up a series of 5 live sessions to cover topics around Decision Architecture. It’s free and exclusive to Uncertainty Project subscribers, but we will have a limited number of spots (just to make sure we can facilitate/manage actual discussion). Sign up for the waitlist if you’re interested!

Upcoming ProductOps talk with Chris Butler

For our subscribers interested in ProductOps, we’re looking forward to this upcoming talk with Chris Butler! - you can RSVP here!