Against Process, For Frameworks

Process, and framework

The two words are used as if they were neighbors, and in some organizations as if they were synonyms. They are neither. Precision here is not pedantry, because the whole argument turns on the difference, and the difference is easiest to see in what each one hands over.

One clarification is needed before the argument can begin, because it is the one most often blurred. In the definition of process, what to do means the actions, not the aim. A framework may be entirely clear about the aim; saying what must be true when the work is finished is not process, it is the minimum courtesy owed to anybody asked to do anything. Process begins when the instruction extends past the aim into the actions and the method, and it is complete when nothing is left for the person to decide.

Ask an organization what it wants from the people it hires for their minds, and the answers are consistent: judgment, initiative, creativity, the ability to handle what nobody anticipated. Then read what it hands them on the first day. A large part of it is a description of what to do and how to do it, written by somebody who is not present, about situations that person imagined, for a reader assumed to need it.

Stated once, plainly: process is an insult to the intelligence it is handed to, whatever the intention behind it, and a framework is the opposite. Not because it is gentler, but because it is the only way judgment has ever been passed from one person to another.

The claim is not that structure is bad, that rules are bad, or that experience should go unrecorded. It is a claim about form. The same experience can be handed over as a fish or as fishing, and the choice tells the recipient exactly what its author thinks of them. The proverb is older than management and better than most of it; the rest of this essay says what it means when the recipient is smart, creative, and hired to be so.

Three questions that separate them

Definitions can be argued with, so it is better to have a test. Any artifact (a document, a method, a policy, a tool, a way of working) can be put to three questions, and the answers sort it. None of them asks what the artifact is called, because the name is the least reliable thing about it: a great deal of process ships under the word framework, and a certain amount of framework is filed as procedure.

Before the questions, the pairs, since the distinction is easiest to feel in things that already exist in both forms:

Turn-by-turn directions

→

A map

A recipe

→

Technique

A phrasebook

→

A grammar

A score, played as written

→

Harmony, and an ear

A script

→

A brief

Orders

→

Intent

A fish

→

Fishing

In every pair the left-hand item answers the situation its author foresaw, and only that one. The right-hand item answers situations its author could not have foreseen, and the price of that is a reader who can think. The recipe produces one dish for a cook who cannot cook; technique produces dinner from whatever is in the kitchen. Nobody asks which is the better gift to a capable adult, because the question answers itself, and it answers itself the same way at work.

The three answers always agree. That is the useful part: in practice you rarely have to ask more than one of them.

Why it lands as one

The word insult is deliberate, and it needs defending, because process is rarely written with any intention to insult. It is written by careful people who want things done properly. The insult is not in the intention; it is in the message, and the message is delivered whether or not anybody meant to send it.

Process has an honest origin, and the origin explains the message. Frederick Taylor's scientific management proposed, in so many words, that the planning of work be removed from the people doing it and given to a separate class of people who would think, so that the doers would only do. This was not a hidden premise; it was the method, and for the pig iron and machine tools he was studying it was at least arguable, since the value of that work lay in motions and motions can be specified. Every process written since carries the premise, whether or not its author has read a word of Taylor: somebody else has done the thinking, and what remains is execution.

Douglas McGregor gave the premise a name when he described two theories of the person that a manager can hold. Under the first, people must be directed and controlled, because left to themselves they will avoid effort and responsibility. Under the second, people exercise self-direction toward aims they have accepted, and seek responsibility rather than avoiding it. Process is the first theory rendered as paperwork. It does not matter that its author privately holds the second. The document holds the first, and the document is what the person receives.

So the message of a process, stripped of its formatting, reads: the thinking about this has been done, by somebody who is not you, and your part is to comply. Delivered to a person hired for the quality of their thinking, that message is an accurate description of an insult. It says that the faculty they were hired for is not wanted here, and it says so in the most durable form available, a written one, to be consulted daily.

It is received accurately, too, which is the part the defenders of process tend to miss. Intelligent people are precisely the people who can read the premise beneath an instruction. They know the difference between being told what must be true at the end and being told which hand to use. They know that a step which exists to remove their judgment exists because their judgment was not trusted, and they draw the conclusion the evidence supports.

Two responses are available to them, and both are expensive. The first is to leave, which is the response of the best of them, and it is why process-heavy organizations tend over time toward the people who do not mind it. The second is to comply, and to stop thinking, since thinking has been ruled out of the job. This is the more expensive response, and it is the more common one, because it is invisible. The organization has paid for a mind and received a pair of hands, and the ledger shows nothing wrong, because the process was followed.

There is a body of research on what constrains creative work, and its consistent finding is not surprising. Teresa Amabile's account is the clearest: creative people can be given clear goals, and often work better for them, but they must be given freedom in the means, and prescribing the means reliably reduces what they produce. This is the insult, measured. Process is the prescription of means by definition, and the reduction in creative output is not a side effect of applying it to creative people. It is what applying it to creative people does.

What is actually handed over

If process is what to do and how to do it, a framework needs a positive description of its own, or the argument is merely against something. Three things are handed over, and all three come from the same place.

All three come from experience, and the word matters, because it is what separates a framework from an opinion. A framework is the residue of problems actually solved, by people who were there, condensed into a form somebody else can pick up. It is the answer to the question every capable newcomer actually asks, which is not what should I do but what did you learn.

This is also why a framework can transmit what a process cannot. Michael Polanyi's observation that we know more than we can tell describes every expert who has ever tried to write a procedure: the procedure captures what can be told, and the part that made them an expert cannot be. Donald Schön found the same from the other side, watching practitioners reason in the situation rather than apply rules to it. A framework does not pretend to contain that skill. It contains the ideas and tools the skill was built from, and hands the building over to the person.

Henry Mintzberg described the alternative at the level of a whole organization. Work can be coordinated by standardizing the process, so that everybody follows the same steps, or by standardizing the skills, so that everybody has been trained into the same body of knowledge and can then be trusted to coordinate by judgment. The first gives what he called the machine bureaucracy; the second the professional one, which is how hospitals, universities and firms of engineers have always worked when they worked well. A framework is the standardization of skill: the form of coordination that assumes a mind on the receiving end.

The clearest institutional case comes from the institution most associated with orders. Armies learned, at cost, that detailed instructions fail at the point of contact, because the situation there is never quite the one the instruction described. The doctrine the United States Army now calls mission command, descended from an older Prussian practice, states the commander's intent and the constraints and leaves the method to the person on the ground, who can see it. It is a framework issued to people in the most dangerous circumstances there are, and it exists because the alternative was tried.

A proverb, read precisely

The proverb everybody knows runs: give a person a fish and you feed them for a day; teach them to fish and you feed them for a lifetime. It is usually attributed to somewhere ancient and far away, and the attribution appears to be decoration. The earliest form anybody has found in English is in a novel of 1885, where a character remarks that if you give a man a fish he is hungry again in an hour, and that if you teach him to catch one you do him a good turn. Her version is the better one, because it names the interval: an hour, not a day and not a lifetime, just the length of one meal.

Read as a statement about work, the proverb says something more exact than it is usually credited with. A process is a fish. It solves today's instance of the problem, it is consumed by being used, and it leaves the person exactly as capable as it found them. Tomorrow's instance needs another one, from the same source, and the source has to still be there. A framework is the fishing. It is slower on the first day, because learning is slower than being handed the answer, and it is faster on every day after, because the person no longer needs the source.

The proverb is quieter about the giver than about the receiver, and the giver is where the interest is. Giving fish keeps the giver necessary. Teaching fishing makes the giver unnecessary, and that is the point of it. Written down, the asymmetry explains a good deal about why process accumulates in organizations and frameworks do not: a process makes its author indispensable, and a framework makes its author obsolete, which is a gift few people are eager to give. It is nonetheless the gift experience owes to the next person, and the withholding of it, however it is formatted, is process.

The two forms also behave differently over time, and in opposite directions. A framework improves in the hands that use it, because its users add what they learned, and it was made of that material to begin with. A process degrades. Each exception becomes a new step and each incident a new clause, and no step is ever removed, so the document thickens until it is consulted only to find out what will be held against whom. Meanwhile the capability of the people underneath it stays where it was, because the document was never designed to raise it.

The shape of it, drawn without numbers, since none are claimed:

The crossover is the whole difficulty. A framework costs more on the first day, which is the day most decisions about form get made, and that is why process is chosen so often by people who would choose otherwise if they looked further along the line. The proverb is not sentimental about the first day. It simply knows about the second.

Why process is chosen anyway

If the argument so far is right, process should be rare among intelligent people, and it is not; it is everywhere. An argument that cannot explain its own counter-evidence is a mood, so here are the forces that produce process, at their real strength. None of them requires anybody to think badly of anybody.

Underneath all four is a fact about knowledge that Friedrich Hayek set out in a different context and that applies here without modification. The knowledge that matters for a decision is largely knowledge of the particular circumstances of time and place, and it exists only where the decision is made. A process is an attempt to make that decision somewhere else, earlier, by somebody who could not have had it. It is wrong at the point of use for the same reason a plan drawn up in a capital is wrong in a village, and it is wrong more often the more particular the circumstances are.

Where process is right, and how to tell

An argument against process everywhere would be wrong, and it would be wrong in a way an intelligent reader would spot at once, since some of the most capable people alive follow checklists for a living. There are three cases in which process is the right form, and the interesting thing about them is that they share a shape.

The shared shape is this. Process is right where judgment is absent, unnecessary, or dangerous to rely on alone: the novice does not yet have it, the machine does not need it, and the checklist guards its known failure. Everywhere else, which is everywhere an intelligent person is doing work that needs one, process is the wrong form, and a framework is the only form that respects what was hired.

Kinds of work, ordered by how much of their value is judgment:

The marker is the point at which the person's judgment becomes the value of the work. Left of it, a capable person can accept a process without any sense of being diminished. Right of it, the same document claims the faculty they were hired for, and the claim is the insult.

Three objections worth taking seriously

Three objections come up whenever this argument is made in public, and none of them is a strawman.

Half of this should be conceded at once, and the test in part one exists precisely because a great deal of what is sold as framework will not survive it. The concession cuts the other way too. A framework can be hardened into a process by the people who adopt it, when its tools become rituals and its ideas become ceremonies performed on schedule; it is then a process, whatever it was designed as. None of this touches the distinction. It shows that the distinction lives in the form and not in the label, which is what the essay claims.

This is right, and it is not an objection, because structure is what a framework provides. The objection confuses structure with instruction, and the two do not even point the same way: one holds the work, the other decides it. Nobody here proposes that a new hire be handed nothing at all. The choice was never between structure and none. It is between a framework and a process, and only one of them respects the person receiving it.

It appears so until one reads what its architect actually did. In Taiichi Ohno's own account, a proper work procedure cannot be written from a desk; it has to be tried and revised, many times, at the place the work is done, by the people doing it, and it exists in order to be improved the moment they find a better way. The pen was in the worker's hand. By the test of this essay that is a framework, whatever it was called on the shop floor: it gained value with a more capable worker, it treated the unforeseen case as the occasion for improvement, and it kept responsibility with the person present. What the imitators tended to copy was the steps, and the pen rarely made the trip with them. That is the difference this essay has been describing all along: the steps alone are a fish, and the pen in the worker's hand is what makes it fishing.

Fish, fishing, and the difference it makes

The difference the proverb names is not the size of the meal. It is who has to still be there tomorrow.

None of this is a purity test. No case is made here against the checklist in the cockpit or the rules given to somebody in their first week. Those are honest uses of process, and what they have in common is that none of them makes any claim on the judgment of the person holding them. The insult starts exactly where that claim starts.

What follows is narrower than the argument. Before writing down what somebody should do, ask what you are actually handing over. If the person on the other end was hired for their mind, write the framework: what you learned, why it worked, what to watch for, what it tends to look like when it goes wrong. Then get out of the way. The rest, they can catch themselves.

Telling one from the other

None of this is worth much until it is pointed at an actual document. Below is the test from part one in portable form, with a fourth question the machines have added since. Pick a policy, a methodology, a playbook, an onboarding guide, anything that tells people how to work, and run it through them in order, answering for what the document does rather than for what its author meant by it.

Then ask the last question of your own writing. If what you have handed the people you hired for their minds would be worth less in better hands, you have given them a fish, and they know it.

This position was formed inside the work rather than outside it. The author's working life has been spent where business, process and technology meet; the description is his own, from 2007, and that meeting point is where processes are designed and where their effect on capable people can be watched at close range. The objection recorded here is the objection of somebody who has watched.

An earlier line of his, from 2015, asked which comes first: the triumphant evolution into a master of deriving solutions, or the sad forgetting of how to study, remember and repeat history. He wrote then that the puzzle applies to the balance between promoting creativity and following what history shows to be proven practice. This essay is one answer to that puzzle. A framework is how history is handed on without forbidding the deriving of solutions. A process is history handed on in a form that forbids it.

The proverb's provenance was checked rather than assumed. Its attribution to an ancient source has no documentary support that anybody has produced, and the earliest English form on record is Anne Isabella Thackeray Ritchie's, in Mrs. Dymond (1885). Part four uses her wording where it can, for the reason given there.

No new data or statistics appear anywhere in this essay, by design. Each source below is cited for the argument it makes, never for a figure, and each one was checked against the record before this went up.