Why the Same Problem Keeps Coming Back at Work

You are not unlucky. You are solving on the wrong layer.

Two layers, not one

There are two layers you can fix a problem on. The first layer asks: what action should we take to make this go away? That is what most of a typical week is spent on. Chris Argyris, who spent fifty years studying how organisations actually learn, called this single-loop learning. You fix the visible thing without touching the rule that produced it.

The second layer asks a different question: what about the way we work caused this to happen in the first place? Argyris called that double-loop learning. You fix the rule, not the symptom.

The reason recurring problems feel inevitable is that almost no one stops to do the second layer. Single-loop solutions are quick, visible, and feel productive. Double-loop solutions are slower, less visible, and usually require admitting that something already in place — a process, a structure, a hiring decision — is producing the problem. So we skip them.

A small example, the long way round

Picture a team that keeps missing client deadlines on a particular type of project. The single-loop response: tighter timesheets, weekly check-ins, a project manager assigned. Sometimes it holds for a quarter. Then a deadline gets missed again. Another check-in is added. Then another. The system gets heavier, and the problem keeps surfacing.

The double-loop question sounds different. Why does this type of project always run late? The answer usually sits somewhere obvious in hindsight. The scoping conversation with the client is rushed because someone is rewarded for closing fast. The team that scopes is not the team that delivers. The delivery team finds out about the deadline two weeks after it was promised. None of those have anything to do with timesheets. They have to do with rules — explicit or unspoken — that govern how work gets shaped before it lands.

You can add timesheets for the rest of your career and the problem will keep coming back. The system is producing it.

Or take a different example, the same shape. A particular role in your business keeps turning over. Every twelve to eighteen months, the person in that seat leaves. Sometimes the exit interview tells you something specific, sometimes it does not. You hire again. You brief the new person more carefully. They settle in. About a year in, the same thing starts to happen — the same fading energy, the same conversations about scope or workload, the same eventual resignation.

Single-loop fix: better hiring, better onboarding, a retention bonus. Sometimes it stretches the cycle by six months. Then the next person leaves.

Double-loop question: what is it about this role, as designed, that no one stays in? Often the answer is that the role sits at the collision point between two parts of the business that have never been properly reconciled, and whoever takes it absorbs that conflict every day until they cannot any more. That is not a hiring problem. No interview process can detect it, and no candidate can solve it once they are in the chair.

What Deming would say

W. Edwards Deming made the same point from a different angle. He estimated that ninety-four percent of an organisation’s results come from the system it operates in, and six percent from the individual effort of the people in it. When the same problem keeps coming back, it is almost always sitting in that ninety-four percent.

This gives you a test. If new people in the same roles produce the same outcome you keep seeing, the problem is not a people problem. It is a system problem. You have been solving a system problem with people solutions, which is why it keeps coming back.

Three signs you are stuck in single-loop

The patterns are unremarkable once you know what to look for.

The fix list grows but the problem list does not shrink. You add another check-in, another report, another step in the process. Next quarter, the underlying problem is still there, and now everyone has more meetings on top of it.

The same name keeps coming up in the post-mortem. If the conversation lands on “well, it is just X being X again” or “Y dropped the ball” — and you have had that conversation about three or four different people in the same role — you are not looking at the problem. The role is producing the behaviour.

The solution is always more. More software, more oversight, more meetings, more hires. Double-loop solutions tend to involve less. Less rework because the scope was honest. Fewer meetings because the right people were already talking. Less management because the system holds.

What it sounds like

Some sentences are reliable indicators that the conversation is staying in the single loop. None of these are bad sentences. They just tell you the team is treating the symptom.

“We need to be more disciplined about [the process].” The discipline is being asked of people who are already operating in a system that produces the problem. Adding discipline is asking them to compensate for a structural issue with personal effort. It lasts as long as personal effort can hold it.

“Let’s make sure this does not happen again.” It will happen again. The system has not changed. The sentence is doing emotional work — closing the discomfort of the conversation — rather than analytic work.

“Once X is sorted, we will be fine.” X here is usually a person, a single project, or a temporary disturbance. Sometimes that is true. More often, X is a symptom of the same underlying loop, and “fine” arrives for a quarter before the next variant shows up.

Listening for these in your own conversations is itself a useful exercise. They are not signs of poor management. They are signs that the meeting is doing what most meetings do — staying in the visible layer because the visible layer is where everyone has practice.

How improvement is actually supposed to work

Deming organised improvement around four steps: Plan, Do, Study, Act — the PDSA cycle. Most teams do two of those four. They Plan (set tasks) and they Do (execute them). The Study step — looking honestly at the gap between what was planned and what happened, and asking what that gap teaches you — gets skipped. The Act step — changing the underlying plan or process based on what you learned — gets skipped with it.

A team that operates on Plan-Do alone will spend years getting busier without getting smarter. The PDSA cycle is double-loop learning given a method. Without the Study, the loop never closes.

What to do this week

You do not fix this by reading an article. But there is a small starter move that opens the right conversation.

Take one problem that has come back at least three times in the last year. Write it on a piece of paper. Draw two columns underneath. In the left column, list every action that has been taken to fix it — every meeting, every new tool, every reassignment. Leave the right column blank for now.

Then, in the right column, write what would have to change for the problem to not be able to happen again. Not the easy thing. The thing you have been quietly avoiding because it would mean re-doing something set up by someone you respect, or admitting an arrangement you championed is not working.

The right column is where the double-loop sits. Most teams do not do this exercise — not because it is hard, but because it is uncomfortable.

Single-loop learning will keep your business running. Double-loop learning is what lets it grow without breaking the same way every year.

FAQs – Frequently Asked Questions