A Field Guide for Switching

Six questions, and three routines for using them

Printable. This page is laid out to fit an A4 sheet without losing its headings, tables or figures.

You have a problem. It is the one in front of you, the one you were already thinking about before you opened this. A student who will not answer. A product that sells badly. A meeting that keeps producing the same outcome. This guide is for that problem. It works out which level of thinking the problem is living at, and it gives you a way to change level when you are wrong. It is designed to be printed and used, and it is deliberately short.

The first thing to say is what this is not. It is not a system to think with. It is a system for noticing which system you are thinking with, and that is a much smaller job and a much more achievable one. Most people can already think well at more than one level. What they cannot do is tell which level they are on, and so they cannot tell when the level has stopped working.

Diagnosis is a cycle, not a staircase. You name the level you think you are on, you ask that level's question, you look at a real case, and then you decide whether to stay or switch. You go round it again, and again. The people who are good at this are not the ones who have climbed highest. They are the ones who come back down most reliably.

The six questions

Each level of the ladder comes with one question. The question is the test for whether you are standing on that level, and asking it is what tells you the level has stopped working. The questions are phrased so that each has a definite answer.

One thing? If you can point at it, photograph it, name the day it happened, it is one case. Ask this when you are not sure whether you are looking at a case or at a pattern. It is the question that catches the person who has one bad afternoon and has decided they have a bad term.

A pattern? Is this the same kind of thing as something that happened before, and how many times have you seen it? A pattern needs several cases before it is a pattern, and the number matters more than it feels like it should. If you have two, you have a coincidence with an opinion attached.

A rule? Is there a rule here that would tell you what to do in a case you have not met? If yes, ask the next question: does the rule have a scope, and are you inside it? A rule that cannot be applied to any case you have not met is not a rule, it is a description.

A structure? What arrangement produces this? Not what went wrong, and not who did it. What the shape of the situation is, who it constrains, where the constraint is, and what crosses the boundary you have drawn. If you can change a part of the arrangement, you are in system thinking.

A loop? Does something come back around? Name the output, name the input it returns as, and say how it gets back there. If you cannot finish that sentence, you have a repeating pattern rather than a loop, and the difference matters, because a pattern can be broken in one place and a loop has to be broken in the shape.

For what? What is this meant to be for, and what is it actually for? These two are usually different, and the gap between them is where the interesting work is. Ask this last, because it is the question that most often makes the other five unnecessary.

Fig. 1 The five step ladder on the left, with the level above marked in accent. On the right, each level paired with its question.

Symptoms

The questions above tell you where you are. The table below tells you where you have got stuck, which is more useful when you already know you are stuck.

You keep collecting detail. You have notes on everything and no decision. You are at the tangible level and have mistaken collecting for working. The next step is not more detail. It is two cases side by side, looking for what they share.

You have a lot of words and no plan. Your notes are full of terms, categories, and labels, and you cannot say what to do on Monday. You are at the abstract level and the categories have become the work. The next step is to write the decision in one sentence, in plain words, with no category words in it.

You are certain and cannot be moved. You have a position that you can defend and that nothing can touch. You are at the principle level and the cases have gone missing. The next step is to find the situation the principle was made for, and check whether it is this one.

You have explained it beautifully and nothing has changed. There is a diagram, or an analysis, or a page of notes, and the situation is exactly as it was. You are at the system level and the explanation has become the activity. The next step is one specific change, made this week, to one part of the arrangement.

You can see it perfectly and there is nothing to be done. You understand the cycle, the constraint, the whole shape, and it feels like fate. You are at the loop level and the description has become an excuse. The next step is to find the part of the loop that is not fixed, and to name it.

You are talking about how to think. The conversation is about frames, models, and which approach is best, and nothing in front of you is being examined. You are at the level above the ladder and the reframing has become a way of not looking. The next step is to name a particular thing and go and look at it.

Fig. 2 Four boxes: name the level, ask its question, look at a case, switch or stay. Arrows run round them.

Here is the guide being used on a real-looking problem, quickly, in the order you would actually do it.

The problem: a small shop is doing well and the owner is exhausted. Everything is fine and nothing is fine. The owner says the problem is that they are working too much, and the team says the problem is that the owner does too much.

Two minutes, name the level. One thing? No, this has been going on for a year. A pattern? Yes, and the pattern is: the shop gets a bit busier, the owner absorbs it, the owner gets exhausted, the shop slows down for a while. The owner calls it a bad week. The pattern question says the bad week is the end of something that started earlier, and that the loop is already visible.

Twenty minutes, ask its question. A rule? Is there a rule that would tell the owner what to do in a week they have not had yet? Perhaps: never work a Saturday without a break in the middle. Hmm, that is a rule, and it can be applied to an unseen week. Does it have a scope, and is the owner inside it? The scope is weeks when the shop is busy and the owner is the only person who can do the thing that only the owner can do. Yes, inside. So the rule fits.

Look at a case. Pick the last bad week and look at it in detail. Monday: owner did the buying, the orders, the two difficult conversations, and the accounts, in that order, with no gaps. Saturday: owner covered because the Saturday person was ill. The case shows something the loop description did not: there was one week where two people were absent, and the owner absorbed both, and the shop survived. The owner did not have to absorb it. The question was never asked.

Switch or stay? The loop analysis was right and it was not enough. It named the shape but not the lever. The case found a lever: the Saturday cover, and who is called when two people are out. So: switch one level down, from the loop to the system, and the change is a named person and a named procedure, made this week, for the case where two people are absent.

Three routines

The routines exist because diagnosis only helps if it happens often enough to catch the problem early. These are not a discipline to be maintained. They are three small habits, and they take a total of rather less than half an hour a week.

The two minute routine. When something starts to feel wrong, and before you start solving it, name the level and ask its question. Two minutes. Write the level and the question at the top of a page. If the question produces an answer, work at that level for a while. If it does not, that is the first evidence that you are on the wrong one, and you have lost two minutes rather than a fortnight.

The twenty minute routine. Once a week, take one thing that has been bothering you and give it a proper twenty minutes at a named level. Not the level you are good at. The level the problem needs. Set a timer, because a routine without an end is a spiral, and the timer is what turns an evening of worrying into twenty minutes of work. At the end, write one line: what changed, what you would do differently, and what it cost.

The weekly review. Once a week, look at the week and ask two questions. Which level did most of my thinking happen on? And where did the week go? The answers are usually the same, and the gap between them is usually the finding. Most weeks are spent at the level you are best at, and most stuck problems are at the level you are worst at.

Fig. 3 A printable sheet. Six questions in two rows, then three routines: two minutes to name the level, twenty minutes to ask its question, and a weekly check.

Using the guide on a decision you have already made

One more use, which is the reason the guide is a field guide and not a theory. The decisions that need it are usually already made, and the thinking that has already happened cannot be undone, but the decision can still be reopened.

Take a decision that was made a while ago: the price, the hire, the thing you stopped doing. Run the six questions over it. The level question will often surprise you, because the decision was made at a level and then defended at a different one. Somebody who set a price on instinct will defend it with a principle about quality. Somebody who made a decision about a person will defend it with a system argument about what the role requires. The gap between the level of the decision and the level of the defence is often the whole problem, and naming it is enough to make the decision available again.

What if you run the six questions and the problem genuinely does not fit any of them? Then the model is not a description of the problem, and you should say so and stop using the guide on it. Some problems are a single thing that has gone wrong once. Some are a mess with several causes that do not form a pattern. Some are a matter of taste, and taste is not a level. The guide is a set of questions to help you notice which level a problem is on, and the honest answer to a problem that fits none of them is that it is not a level problem.

Now change the condition: what if the problem fits several levels at once, which is the common case for anything that has been going on for years? Then start at the lowest level that has anything to say, and work up one step at a time. Do not start at the top, because the top will always have something to say and the top will always sound convincing, and the point of the guide is to find the level where the answer is not yet available.

"The level is a property of the problem." It is not, quite. The level is a property of the way the problem is being looked at, and most real problems can be looked at from more than one. A student who will not answer can be looked at as a particular behaviour, as a pattern across the class, as a rule about who speaks, as a structure of who sits where, or as a loop of quiet answers leading to more quiet answers. All five are true, all five are about the same student, and they call for different actions.

The other misconception: "the guide is for when you are stuck." It is for when you are not stuck and do not know that you are about to be. The two minute routine is a check, and checks are cheapest before the problem has grown. Running the questions on a small problem takes two minutes and on a large one takes a season, and the difference is entirely in when you started.

Limits

The guide is a set of questions, and questions have a failure mode of their own: they can be asked without the answers meaning anything. It is entirely possible to run all six on a problem and to write down six sentences and to have learned nothing, because you answered them at the level of the question rather than the level of the thing. The check for that is the twenty minute routine, which is about looking rather than about naming.

Second limit: the guide describes a position, not a method. Naming the level does not tell you what to do at that level, and some levels are hard to work at. It is easy to identify that you are at the system level and hard to actually draw a good system diagram, and the guide will not help with the second thing.

Third limit: the levels are the author's own model, and the guide is calibrated to it. A reader who has a different and better model should use the questions, not the levels. The questions are the portable part.

Stuck

You are stuck with the guide when you use it to describe problems rather than to change them. The tell is that your notes are full of level names and full of nothing else. You have renamed the problem, which felt like progress, and the problem is exactly where it was.

The way out is to make the guide produce an action. The final question in the cycle is switch or stay, and it has to have an answer in the form of a thing somebody will do on a particular day. If the cycle ends with an observation rather than an action, you have stopped one step short, every time.

Sources

The five levels, the six questions, and the three routines are the author's own. No specific framework or study is cited.
The shop example is hypothetical. No real business, price, or figure is used.
The one page summary figure is intended to be printed, and the field guide page is laid out to fit an A4 sheet.