The Challenge of Complexity

Image for post
Image for post

Complexity is a term familiar to individuals, foreign to organizations. Understanding it and its implications can transform everything we do.

We are an organization devoted to applying best practices” is a variant of an expression I often hear in my work, particularly within healthcare. While those who utter such phrases are usually not intending to deceive (me or themselves), they are really telling a lie or are showing that they don’t know where they are going. The reason has not to do with their intent, but the complexity in which they operate.

Complexity is a term that is familiar to us as individuals but remarkably challenging for organizations. The reasons have much to do with trust, change, and the risks that come with not having full control over the context of our work and the outcomes associated with it.

The seductive nature of the “best practice” and the prescription for change in 5,7, 10, 12 or whatever easy steps are something that is endemic in our society. These forms of thought suggest a linear trajectory of events, suggest an ability to control for externalities and parse out their impact and provide a prescriptive solution that removes much of the worry about unknowns. But H. L. Mencken’s often quoted phrase (which I’ve used often) suggests the folly in this.

When “Best Practice” Isn’t So

‘Best practice’ is all about applying the best evidence consistently toward a problem and, as the name suggests, doing the best of all alternatives. That sounds reasonable enough, yet the foundation for this evidence is based on a model that is largely inconsistent with complexity.

Organizations thrive on the ability to create coherence and coordination between people, activities, and outcomes in pursuit of goals. The more people, activities, and desired outcomes involved, the more the effort is placed on control to generate this coherence and foster coordination across all these variables. This all takes place within a system and how we understand the role of variability and patterns of energy within these systems is what makes the idea of ‘best practice’ a problem in healthcare and most human services.

Best practices work when the system complexity is low or non-existent. These are sometimes called simple systems and are the kind found in assembly plants or basic electrical or plumbing systems. Such systems enable us to learn from experience and predict with high confidence how future actions will unfold. These allow for standardization, automation, and predictive models that truly enable us to create ‘best practice’ because we can test alternatives and determine what ‘best’ is because we have the measures, metrics, and methods to assess that.

Complicated systems are those with some complexity, but enough stability that there is an ability to derive the type of evidence that might require some expert interpretation, but allow us to assess better alternatives. We might not always have the best practice, but we certainly can determine better practices. Large-scale mechanical or linear systems with clear outcomes (e.g. flying an airplane, organizing a conference presentation) fit into this category.

In both of these systems, we can plan using available evidence to make strong assertions about possible outcomes and have some confidence that our actions will yield those outcomes or something close. Traditional strategy or planning models work reasonably well within these systems. Healthcare and human service systems have many areas where this kind of thinking is appropriate and reasonable, but also many areas (including the enterprise as a whole) where it’s not.

This is where we find ourselves with hammers looking for nails.

Accounting for complexity

Image for post
Image for post

Just as no one (at least no one I’ve met) would consider drawing up a flowchart and showing a prospective mate the planned trajectory of their dating relationship with milestone targets and deliverables, no organization should think that they can pick some arbitrary target (e.g., ‘5-year plan’) and impose controls and other structures to force outcomes in complex systems. Like dating, much of human service work is complex.

To design organizations to meet complexity, we first need to consider how we are primed to resist complexity in our organizations.

1. Our education system is designed for linear, progressive modes of learning, not discovery and non-linearity. We sit kids (and adults) in rows, we talk at them, we present material front-to-back. In short, we don’t design education for learning, but for information transmission. Complexity is all about learning. Every situation has a degree of novelty to it that presents new challenges and what happens today might not be the same thing that happens tomorrow even if much is similar. Teaching to discover, adapt, play and risk is something our system doesn’t do well. How can we expect complexity and systems thinking to thrive when the muscles used aren’t developed?

2. It’s more convenient to think in dichotomies than spectrums. As I’ve written previously, spectral thinking is something critical to many of the issues we face in complex systems. Good/bad, strong/weak, X/Y lose their meaning in complex environments where the dynamic context makes these assessments not useful. Utility — the ability to take something and use it for a beneficial purpose — is the quality marker for learning. (For example, burning your hand on a hot stove is a useful learning outcome for working in a kitchen, not good or bad overall).

Of all the dichotomies that work in complex systems, only Ying/Yang comes close. But its a more difficult concept to grasp that maybe things aren’t all one way or the other, that there is use in even something that isn’t well constructed. This problem (and the ones that follow) are tied to the first one: education and learning systems are not set up for this. We are primed for either/or thinking.

Dichotomous thinking also allows us to generate targets and success metrics even if these might actually end up hurting us in the end.

3. Our decision-making tools are ill-equipped to handle ambiguity. Healthcare is a great example of how badly we are at complexity thinking. Consider the systematic review, often viewed as the gold standard for evidence for adoption into healthcare organizations. If something (procedure, task, idea) has some support from a systematic review, then one would imagine that it would be a likely candidate for adoption into wider practice within the system. Right? No.

Surprisingly, even systematic reviews of systematic review use show a mixed bag in the eventual adoption of evidence into practice. Systematic reviews are designed to reduce ambiguity, but really illustrate how much there is within the wider system. A systematic review looks only at the evidence created (and published), it doesn’t include all those questions that were never asked, never funded for inquiry, or couldn’t be structured into a study that meets the strict criteria for inclusion in a good review.

Systematic reviews are, by design, reductionistic in their approach to complexity. What allows us to do systematic reviews is standardization and consistency and that is one of the very things that complexity fowls up.

4. Our institutions are resistant to complexity. Complexity takes time, nuance, and relationship development; all the things that screw up plans and timelines created by organizations. You can’t plan a relationship, but you can anticipate some things. You might even be able to use scenario tools and strategic foresight methods to anticipate what might happen, but you can’t plan it.

John Lennon was right:

Life is what happens when you’re busy making other plans

While we plan, the complex systems move along. We can plan and fail, fail and plan, or plan to fail and work build the strategic foresight to know what to do with these “failures”.

Designing organizations for complexity

Image for post
Image for post

So now what? Being aware of these things is a start, but creating organizations that are designed for complexity is what will create lasting change. Where do we begin?

Like with many complex systems, some small changes done well can have large resulting effects.

  1. Learn to recognize the difference between a simple, complicated, complex and chaotic system and the means to identify when those systems present themselves. The Cynefin Framework is a tool/method that can support this work. Matching the level of complexity of the situation with the kinds of evidence, approaches, and tools that will support strategic action in those situations, organizations can best align their resources with reality, not preconceived ideas.
  2. Complexity poses a challenge for evidence as there is almost always a mismatch between what is learned from one context to another. This can’t be addressed by simply doing more analysis of data, rather it requires something social: sensemaking. Sensemaking is a social activity that brings insight from practice to the data on hand to ask and answer: what does this mean for us? It recognizes that data is gathered in a context and the decisions arising from that are also applied to a context. Creating space to gather the right data and evidence and then connect it to context is key.
  3. The social aspect of sensemaking means creating a culture where people can listen, learn, speak, and share their thoughts on what they do. This means finding like minds, sharing stories, and building networks within an organization to generate the requisite variety (diversity) needed to make sense in complex systems. It means creating space for fostering relationships between those in different sectors, divisions, roles, and sites within an organization. The more that people can comfortably work across boundaries — even temporarily — the greater the insight that will be generated. For example, Freshbooks encourages employees to go on ‘blind dates’ to connect people in different parts of the organization to help them learn about what each person/job does.
  4. Pay more attention. It’s remarkable how many organizations sleep-walk through their planning process generating 5-year plans and other linear evaluation systems that are bad-mouthed by staff before they even get launched. This is as much about adhering to traditional models as it is a failure to pay attention. Too often organizations are busy on the next 5-year plan (which is largely arbitrary and has little connection to strategy) before fully understanding what happened in the current plan.
  5. Setting up a mindset for developmental evaluation is another way to change the way your organization views innovation, adaptation, and complexity. Developmental evaluation is an approach to gathering data on programs in context and linking that data to strategic considerations and design. A shift in mindset is difficult, but also simple and can enable your organization to change its relationship with data, evidence, and decision-making to better work with complexity, rather than challenge it.

These are starting points, but not all of them. Addressing the challenge of complexity is, ironically or perhaps appropriately, complex. But the challenge of dealing with the negative outcomes resulting from overly simple approaches to dealing with complexity will ultimately be far more so.

Image credits by Cameron Norman and by Alex on Unsplash

Designer, evaluator, educator, & psychologist supporting people in making positive change, by design. Principal @censeltd @censeacademy

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store