Unraveling Uncertainty and Complexity

Uncertainty is found when events and outcomes are unpredictable, like when the causes and effects across an organization are not well understood.

Complexity is found when there is a multiplicity of elements that are strongly interconnected. The greater the number of relationships between elements (e.g. items and/or people), the more difficult it is to understand.

In complex systems, a change in one place can (and usually will) cause unintended changes in other places. Cause and effect get obscured by the many interconnections, and it will not always be clear which factors are important in the decision-making process.

So greater complexity will drive more uncertainty into decision making. And we often respond by introducing more complexity in our organization to address the uncertainty. Help!

We can start by asking: What are some of the sources of this complexity in organizations?

First, we acknowledge that our modern organizations are complex adaptive systems. Then we model them as (1) a set of autonomous agents, (2) that are interconnected, where (3) the system behavior emerges from the independent decisions made by the agents.

‍

‍

Natty Gur introduced a map for exploring the sources of complexity in an organizational system like this. In it, he advocates for studying three factors within the system:

1. Competing Drivers
2. External Noise
3. Internal Events

[Credit: “3 simple steps to find the causes of complexity and reduce uncertainty”, Natty Gur]

Let’s walk through this kind of exploration, then consider how we can apply techniques for navigating uncertainty to manage the spread of complexity.

Competing Drivers

‍

‍

Drivers are the things that influence your behavior, as an agent in the system. In our organizations, these show up as things like productivity, quality, velocity, agility, and profitability, just to give a few examples. These can be the things that define your role, the changes you seek to improve, the corporate goals you support, and/or the fires that you are currently trying to put out. Drivers compete with each other to guide agents to a preferred behavior (for the driver), and together, they can push an agent in different directions. Some pairs of drivers might even force trade-offs.

Multiple competing drivers create complexity in the eyes of an agent. And as different agents respond to these drivers in different ways, the complexity of the system can grow, as a response.

External Noise

‍

‍

External noise refers to the amount and frequency of influence coming from the external environment of the system. This is the cacophony of the world around you, relative to your system boundaries.

While there will always be external noise, when the amount and frequency grows, our responses will often add to internal complexity. Gur says that external noise “shakes the system and increases the interconnectivity between agents” which leads to emergent behavior that could be good, or bad.

More importantly, this “shake to the system” can (and should) be a driver for continuous change, since the status quo is rarely robust enough to handle the “shakes”.

Internal Events

‍

‍

As agents respond to drivers and external noise, their actions create internal events that produce emergent behaviors and (possible) momentum for adaptation.

Is a software product receiving a lot of bug reports? Then that quality driver is going to influence how you handle that increased amount and frequency of noise. You prioritize the bug fixes (for a while), which is now an internal event.

But in a complex system, an event in one agent will ripple through to other agents, for better or for worse. Observing this system's behavior is all you can do. Leaders in the system can then seek to tweak the environment, perhaps by recalibrating the drivers. Gur notes that,

"Compared to the previous two conditions, organizations have more control over the way they react or respond to internal events."

So this is a model of where complexity comes from in an organization. The complex system itself is one we barely understand - there is significant uncertainty in how competing drivers, external noise, and internal events yield system behaviors.

How can we start to navigate this uncertainty to drive learning? Here are three places where our techniques can help us expand the model.

Capture and Share Beliefs

‍

‍

The external world around us is noisy. We try to make sense of it, but we are forced to work with incomplete information and make many assumptions. We rely on our individual experience and knowledge.

If we work together to document and share our interpretations of external noise as beliefs, we can build a shared understanding that draws on a group’s experience and knowledge. And the dialog around these outside-in perspectives is critical to calibrating the drivers to set a strategic path.

Prioritize with Even-Over Statements

‍

‍

When there are multiple competing drivers, autonomous agents are left to set their own priorities across them. This is how organizations become misaligned, and why they add complexity to solve this problem.

Leaders accountable for the behaviors of the system (and the outcomes produced) should clarify the relative importance of the drivers. This can be done by acknowledging that while both (or all) drivers are important, at this point in time, the leaders believe that agents should let their behaviors be more influenced by Driver 1 “even-over” Driver 2.

Take the Outside View

‍

‍

When agents are left alone to treat their problems as unique, they can fall prey to biases that steer them astray (i.e. confirmation bias, loss aversion, etc.). Instead, leaders of the system can encourage them to “take an outside view” and look for others that have faced similar problems. In organizations, that often involves studying how customers have been served by competitors, or even have solved similar problems themselves.

Organizations large and small have found a strategic advantage by introducing a driver for customer needs. Customer-centricity is an outside view that can reduce complexity across the system through better alignment of agents. It can also work to reduce uncertainty in the external noise coming from customers. Reducing uncertainty (in an organization’s knowledge of customers) while reducing complexity inside the system is a powerful strategic lever.

‍