She was in a noticeably better mood. It wasn't the type of mood improvement made possible by a bit of good news. It was bigger. It was deeper. It may have been the cheeriest I'd seen her in the fifty-or-so interactions we've had.
She'd been cutting my hair for more than five years and so every five weeks, as part of the regular 30-minute chit-chat catch-up, I got an update on how her business was doing.
In a word: Better.
A few months back she had made two changes that were starting to bear fruit. First, she decided instead of working six days a week she was going to work four and steer clients to her availability as opposed to the other way around. Second, she increased her prices.
More cash in her pocket. More time for her art. And more time for herself. After years and years of single-loop learning, struggling with the cash flow challenges of a sole-proprietor and in constant burnout mode, she dabbled in a bit of double-loop learning. And it worked!
Double-loop learning is a funny name for the learning we do when we move beyond just solving problems (single-loop learning) and explore whatever it is we're trying to do more holistically. Holistically is a funny word but I believe its collective specificity and vagueness capture what double-loop learning is all about: the product of inquiry, reflection, experience, and figuring it out.
The verbs of those nouns make double-loop learning sound good. It is. And it should be how we work much more often than we do.
But most people generally aren't very good at double-loop learning. We often don't question underlying assumptions, norms, and objectives in most situations, and we really should.
There are different contributing factors for why that's the case and the first one worth addressing is because we don't know double-loop is something we should be doing.
Instead we've learned to work like thermostats: see problem, solve problem. Chris Argyris shares his canonical analogy:
... a thermostat that automatically turns on the heat whenever the temperature in a room drops below 68 degrees is a good example of single-loop learning. A thermostat that could ask, "Why am I set at 68 degrees?" and then explore whether or not some other temperature might more economically achieve the goal of heating the room would be engaging in double-loop learning.
Single-loop learning is problem solving. And our jobs do require problem solving. So it's not that single-loop learning is something we shouldn't be doing.
It's that straight line problem solving often isn't enough. When it comes to working in complex operating environments, simply solving the problem without appropriate consideration of the problem's context can 1) prove fruitless, 2) produce an unsatisfactory result, or 3) make the problem even worse.
See Problem, Solve Problem?
Here's an example from a healthcare contact center struggling with a panoply of challenges. Those challenges manifest to the outside world as long caller wait times—"We're currently assisting other callers. Thank you for your patience."—with demands from the environment (patients, provider offices, and executive leadership) to reduce them.
The long caller wait time problem has been ongoing and is very public internally and so it receives a lot of attention when it requires attention. When that attention is given, caller wait times are reduced, and the environmental demands for improvement go away. That is, until the problem again becomes something everyone is complaining about—the hot and cold cycle sounds a bit like a thermostat doesn't it?
Along the way someone came up with a metric, that average caller wait time should be less than 60 seconds, which immediately became a goal, and since then the (new) management team hasn't had to wait for protests from the environment before deploying interventions in an attempt to meet the goal.
See problem. Solve problem. See problem. Solve problem. See problem. Solve problem.
But no one asked—whether they thought to, weren't encouraged to, or declined to—what they should have been asking: Why are these interventions, the same interventions we've been trying time and again, not working?
Double-loop learning would help this contact center in two ways:
- Informing and trying new interventions by learning what worked and what didn't from previous attempts; and
- Instead of, for example, only asking "How do we decrease caller wait times?," a double-loop learning approach would explore why the problem exists, the context of the wait-time problem, what's contributing to it, different ways to think about it, and alternative ways to approach it.
Double-loop learning is learning. It's reflection on the way you think. It's, as this Farnam Street blog post states, "the key to turning experience into improvements, information into action, and conversations into progress."
It may even sound like how you work already, but I have bad news: Chris Argyris's research says you probably don't.
Consistently inconsistent
There are other factors aside from our lack of awareness that get in the way of double-loop learning including workplace culture, full calendars, an achingly long to-do list, others to be sure, and most importantly: our Big Egos.
Argyris wrote an entire article about this, provocatively titled: "Teaching Smart People How to Learn." His central premise: because smart people (you) are generally successful, they tend not to be as open to learning, and worse: just knowing this information isn't enough to overcome the challenge.
"Professionals embody the learning dilemma," writes Argyris, "they are enthusiastic about continuous improvement—and often the biggest obstacle to its success."
The reason is because we have a defensive posture toward reflecting on our individual contributions to any situation. We don't want to admit, often to ourselves and more often to those around us, that we were wrong or we failed. And when the working environment doesn't promote the admission of errors, the issue is worse, and gets worse as we climb the hierarchy and experience more success.
It's easy for us to point at external factors and difficult to turn inward when reflection is needed, as it often is. It's the difference between what Argyris calls our "theory of action" and our "theory in use."
"It is impossible to reason anew in every situation," writes Argyris, "If we had to think through all the possible responses every time someone asked, 'How are you?' the world would pass us by." So instead of constant reasoning, we create shortcuts for why things are the way they are.
"Therefore, everyone develops a theory of action—a set of rules that individuals use to design and implement their own behavior as well as to understand the behavior of others," he continues, "Usually, these theories of actions become so taken for granted that people don't even realize they are using them."
A paradox of human behavior is we believe we're employing a theory of action in any given situation, but if we were to look critically at our actual behavior, we'd discover our defensive posture and a different theory in use.
"Put simply," he writes, "People consistently act inconsistently, unaware of the contradiction between their espoused theory and their theory-in-use, between the way they think they are acting and the way they really act."
To summarize: There is a discrepancy between what we think guides our actions and what our actions actually are. We'll tell everyone, including ourselves, we're interested in learning and improvement (theory of action), but as a result of desiring to avoid embarrassment or threat and feeling vulnerable or incompetent when the focus of learning and improvement turns to us, we generally attempt to avoid learning and improvement (theory in use).
And this isn't just an individual problem. It's also a gigantic problem for any organization.
Considering how we work
That's because single-loop learning is how organizations are managed, too.
In a simplification to be sure, many of our organizations and some of the bosses we work for are analogous to the thermostat recognizing performance is cooling and, as a result, turning up the heat, never stopping to consider that how we work, manage, and organize may be part of the problem—if not the problem.
We—and this is the royal We—rarely reflect on how we do our work. We just don't.
Here is an example illustrating what I mean: We've all participated in a process improvement event that led to better outcomes—but did you consider, or know anyone to consider, whether the method used to organize the process improvement event was the best method?
It just happened, right? Everyone just accepted the approach to the process improvement event without a minute of conversation on the proposed method.
The outcome was probably acceptable—but that isn't the point. Here's the point: What if the outcome could have been even better if the problem had been approached more appropriately?
Why is it we so rarely consider whether how we're working is actually the best way to work? Have you ever considered—or known anyone to consider—the methods used to produce the work? Why not do more double-loop learning on the work itself?
Let's look at meetings. How many miserable meetings do you attend?
The next time you're in a miserable meeting ask yourself: How might this be better? Or:
- What would happen if this meeting didn't exist?
- Why does this problem require a meeting like this?
- What behaviors does this bad meeting promote?
- What work does this bad meeting produce?
You don't have to stop there because there are many more questions worth asking. And not only about meetings.
Take the annual review process as another example. Does it work? Is it useful? Does individual goal setting lead to what we're desiring? Are there better ways to manage performance? Is performance effectively managed using an annual goal setting and review process?
Work doesn't have to be the way it is just because it is the way it is. It's okay to question our methods of management and organization. In fact, we need to start questioning our methods of management and organization because how we do the work dictates the work we get.
"If you've spent your working life in a command and control environment you'll assume there's no other way to manage," writes John Seddon, "It then is logical for you to improve results by just being better at 'command and control.'"
Double-loop learning gives us the opportunity to revisit our misinformed logic and explore a different, even better, way to work.
Moving beyond best practice
For two years early in my career I sat in on the weekly—get organized, discuss what needs discussing, make collaborative decisions—executive meeting. It seemed no matter the issue, there was one executive in particular who always asked some variation of the question, "What is best practice?," as if the practice of asking the question resulted in a solved problem.
I was always bothered by it and never knew why. Now I do. It was misinformed theory.
While there are still problems that benefit from the best practice treatment, there are far fewer of them than any of us might expect, and fewer with each passing year.
Best practice problem solving is perfect for complicated situations. But we work in complexity. And command and control management, the way most of our organizations are managed, was designed for an environment which only produces complicated problems.
We need to work differently in an environment that produces complex problems because complexity creates new problems which have never been solved before. New problems can't be solved by a best practice solution because a best practice solution doesn't yet exist!
Complex problems can only be managed using an approach that starts with recognizing the problem's context.
Attempting to solve a new problem with a best practice solution often results in failure: including trivial failures like bad meetings, useless performance appraisals, and even less-than-optimal process improvement outcomes.
And even trivial failure can be, and should be, a trigger to instigate double-loop learning. "If a new idea doesn't work," states that Farnam Street blog post, "it's time to try something else."
Double-loop learning gives us the opportunity to try something else. Because what double-loop learning allows, instead of zeroing in on a problem just to solve it, is to explore why a problem exists.
It gives us a way to explore a problem's context. It gives us an approach to solve problems in complexity because it provides a method to learn. And our jobs increasingly require learning about the problem we're solving as we're solving the problem.
Our jobs require double-loop learning.
