Not Continuing Can Be Reread as the Normal Mechanism at Work

You start writing in the form you settled on and stop in the third week. The plan to secure the morning hours returns to the old way of getting up within a fortnight. You try a new way of presenting the work and it ends after two attempts with thin responses. The same thing has happened repeatedly, and each time it gets filed under your own character.

What this article hands over is not a management method for keeping things going. The single phrase it does not last points at three separate phenomena. A form where nothing gets started, a form that starts and stops early, and a form that continues while changing into something else. The three differ in cause and in prescription, and they get handled under one phrase.

The advertised prescriptions include ones that genuinely work. Fix the time and the place. Shrink the unit. Reduce the effort needed to begin. All six work, and declining to solve it with willpower does not conflict with the measurements examined below.

Where they land, though, is almost entirely on the side of starting. They connect for somebody who is not getting started, and for somebody stalling after starting there is a different place to aim. Choose a prescription while the phrase stays undivided, and the work proceeds without settling which of the three it is aimed at.

The point of stopping also gets discussed in days. Three days and out; twenty-one days to a habit. Look at the measurements and days alone settle little. The period to establishment spans a fourteenfold range, and skipping a single day did not much change the trajectory. What moves whether things stop reads as the conditions that were supporting the action more than the number of days.

On the starting side there is a further dividing line that has nothing to do with size. The same person starts work that was requested and does not start work they set themselves. What differs is not the volume of work but whether the finished state has been settled.

There is also an asymmetry of records. Only the side that continued leaves anything. How many times and at which point things stopped is, by default, not to hand, so the material for changing the next design goes missing along with it. That it remains on one side only is not a matter of attention: output occurs only on the side that continued.

Set out below, in this order: the basis for separating the three, where the advertised prescriptions land, that stopping happens when conditions disappear, the measured period to establishment, the completion condition that governs starting, the handling of the pushing-back force, and how to keep a record of stopping. The mechanism of that force itself is handled by Mental blocks are not a defect, and the binding of habit to context by Change where you return to before changing the habit.

What Is Actually Ending When You Say It Does Not Last

Lay out the states described as not lasting and at least three separate forms appear.

The first is a form where nothing gets started. You have decided to do it, and when the day arrives you do not put a hand to it. Nothing has begun, so nothing lasts or fails to last.

The second is a form that starts and stops early. It runs for days or weeks and then breaks off. This form is the centre of what is called not lasting.

The third is a form that continues while changing. The frequency drops, the quality drops, it becomes something other than what it was for. Outwardly it continues, so it is not recognised as having stopped.

The three differ in cause and prescription. The first is a problem of starting, the second of holding, the third of purpose. Different stages have stalled, and one phrase covers them, so the prescriptions mix.

None of them lasted, so the sorting makes no difference. It can be read that way. Even so, aiming a response at the wrong one changes the result. Fixing the time and the place does not work at a stage where nothing is being started, because the execution that would be fixed has not occurred yet. So what happens when shrinking the unit is aimed at the shape that stops partway? Starting gets lighter, so the run of days gets a little longer. It then stops at the far end of that run, which puts the stopping point a few days later.

The three can also run at the same time. Starting is late, what starts is short, and what continued changed shape. Fixing any one of them then does little to change the whole. There is an order, though. Holding cannot be discussed while nothing starts, so the first is examined first.

Which of the three is in play can be settled from the week it stopped. For that week, can you recall what was in the schedule? If you can, it is likely the second form, stopped by a change of conditions. If nothing comes back and it simply disappeared at some point, it is the third form, or a first form where the start was shallow.

The proportions shift by phase. Early after going independent the first tends to be common; once a form has settled the third tends to increase. A change of shape preserves the outward form, so it is found last. What usually prompts the noticing is a thinning on the side of results.

The division also fails in one case: where the form itself did not suit the conditions of that period. Then starting, holding and changing all issue from the same cause. What distinguishes it is whether there was ever a stretch in which the form was sustained. If there was never one, suspect the form first.

And the phrase the type who cannot keep things up folds the three into one while placing the cause in disposition. Placed there, the prescription becomes correcting the person. The more experience accumulates of correction failing, the more plausible the disposition account becomes. An account with no picture of how it would look if wrong lets no counter-evidence in from anywhere.

The Advertised Responses Lower the Barrier to Starting

All six of the advertised responses are sound, and where they act sits on the side of lowering the barrier to starting.

Fix the time and place. Add habits one at a time. Set the first target absurdly small. Record progress and results so they are visible. Reduce the effort required to begin. Solve it with the design of behaviour rather than the quantity of will.

As prescriptions, all six are sound. The conclusion that this is not a matter of will is also consistent with the measurement examined below.

Where they work splits in two. Fixing time and place, reducing effort, small targets. All of these lower the barrier to starting. Recording and visibility supply a return immediately after execution.

So the six land on the starting side and the immediately-after side. That is a correct design. What they do not reach is the pushing-back that operates on change itself.

The expectation that a lower barrier will beat the pushing-back holds in some cases. Where the change is weak in that respect — short in duration, barely altering the shape of daily life — lowering the barrier is enough for it to settle. Many habits do settle this way.

The pushing-back tends to grow stronger as the change grows larger. And changes that alter the shape of a business — changing what you handle, changing the price, changing who you address — touch both daily life and self-description. Lowering the barrier here does little to lower the other side.

The six also include combinations that sit badly together. Recording and visibility create an occasion for assessment at every execution. That occasion, when execution lapses, displays the lapse in visible form. While things continue it supplies motivation; after a break it becomes a barrier to resuming. The same prescription runs in opposite directions depending on the phase.

Place the position of the prescription in one line. Systematising is not a tool for beating the pushing-back but a tool for making it unnecessary to beat. With the barrier low enough, execution is over before the force gets to work. So the larger the change, the further it falls outside this tool’s range.

Of the six, recording and visibility do reach beyond starting, since they leave behind what happened after execution. What is left, though, is a record of the executed side. The side that stopped is not in the scope of the record.

And when the system fails to work, the conclusion returns to the person. With the prescription side known to be correct, the only explanation left sits on the executor’s side. Lowered this far and still not lasting, so it must be the disposition. The more correct the prescription, the more plausible the wrong cause looks.

“Three Days and Out” Is a Phrase About Duration, and Stopping Happens When Conditions Disappear

There is a common phrase about the number of days to stopping: three days and out.

The phrase expresses not-lasting as duration. Ending within a short period is the problem, in that framing. And with duration as the frame, the prescription becomes extending the duration. Keep going a little longer.

What actually happens is not necessarily a matter of duration. Things stop, in most cases, not because days have elapsed but because conditions changed.

A form built around weekdays breaks at the weekend. It breaks in a week with travel or visitors. It breaks in the busy season. In each case it is not the days but the disappearance of the conditions that were supporting execution.

Once broken, resuming becomes less likely, because execution was bound to the conditions. When the conditions return, it ought to resume automatically, and the interval that intervened creates a perception that the thing is over.

Does skipping one day end it? No. In the measurement of days to establishment, skipping a day did not much affect the trajectory. Breaking off is not itself a big problem. What becomes the problem is a break after which the supporting conditions are not restored.

Conditions changing cannot be avoided. The shape of daily life moves, and business has its busy and quiet periods. So rather than keeping conditions from disappearing, deciding in advance what happens when they do is the realistic form. What divides the cases is whether a restoration procedure is held.

The phrase carries something else as well. It contains an assessment of the person who stopped. It is not a description of the fact that something did not continue but carries the implication of somebody who cannot keep things up. Describe your own state in a phrase carrying that implication, and a verdict is issued along with the description. Once a verdict is issued, the step of investigating the cause is skipped.

Ending quickly is also not always the problem. You tried it, found it did not suit, and stopped. That is a judgement, not a failure to last. The same observation — over in three days — can be the result of a judgement or the result of pushing-back. What distinguishes them is whether you can say what became clear at the point of stopping.

Handled as duration, the cause of stopping goes unobserved. Over in three days, over in a fortnight contain nothing about what happened. What the record holds is only the number of days, and days yield nothing about what to change next. You try the same form again and stop at the same place.

The Period to Establishment Spans a Fourteenfold Range

The period to establishment carries, alongside a median, a fourteenfold range. The received accounts drop that range and circulate a single figure.

A 2010 tracking study by Phillippa Lally and colleagues at University College London was built so that 96 participants each performed one behaviour relating to food, drink or exercise daily, in the same context, over twelve weeks (Lally, van Jaarsveld, Potts & Wardle, 2010, European Journal of Social Psychology 40(6)). What was tracked is how automatic that behaviour became.

Two results can be drawn out. First, reaching 95 per cent of the automaticity ceiling took 66 days at the median. The individual range, however, ran from 18 to 254 days, nearly fourteenfold. Second, skipping a single day did not much affect the trajectory of growth in automaticity.

So keep going for 66 days? It does not read that way. It is a median, and the two ends sit fourteenfold apart. It is not a figure usable as a single benchmark. And the commonly repeated claim that twenty-one days makes a habit is not supported by this study. Where that claim spread from lies outside the study’s scope. What is certain is one point: the tracking side does not support the figure.

The behaviours handled were everyday ones — food, drink, exercise. High frequency, short duration each time. Days cannot be carried over as they stand to business tasks like publishing or pricing, which are low in frequency and long in each occurrence.

The measurement also carries design implications. Participants selected one behaviour and performed it daily. Selecting one, and daily, are built into the premises. About weekly work, or about starting several things at once, this figure says nothing. The reach of a measurement is exactly as wide as its design.

How the binding of behaviour to context works is handled directly by Change where you return to before changing the habit, so only the side of duration is examined here.

The wide circulation of a benchmark in days has a side effect: falling short of it gets recorded as failure. Somebody who finished in three weeks holds a record of not reaching the received figure. Knowing the measured range, that is merely a value inside the distribution. A benchmark that hands over a single number is what manufactures the record of failure.

What Governs Starting Is Not Size but Whether a Completion Condition Exists

What governs whether starting happens is not the size of the work but whether a finished state has been settled. Most of what is called procrastination falls into the first of the three forms, the one where nothing gets started.

It is usually explained as a shortage of will or discipline. Lay out the work that does get started beside the work that does not, though, and a different line appears. Work that gets started has an image of the finished state. Work that does not has none.

Concretely: requested work tends to get started, because what to do and how far, for it to be complete, has been supplied from outside. Work you set yourself tends not to get started, because the condition for completion has to be settled by you, and while it is unsettled the first move is unsettled too.

So the barrier to starting is not the size of the work but the absence of a completion condition. Shrink it, and without a completion condition it still tends not to get started; enlarge it, and with a clear completion condition it tends to.

The line is consistent with research on the goal side. The paper in which Edwin Locke and Gary Latham summarised thirty-five years of work shows that specific and difficult goals lead to higher performance than vague do your best goals or no goal at all (Locke & Latham, 2002, American Psychologist 57(9)). What is working is not only the height of the goal. Specificity is what settles what to do and how far.

A do your best goal is weak not because motivation is short. It is that where the end lies has not been settled, so the level of performance cannot be judged by the person doing it. Much of the work that cannot be started has this shape. Strengthen the publishing. Review the product. Neither defines a finished state.

Does starting small not settle the completion condition by itself? Sometimes it does. Thirty minutes only is an operation that installs time as the completion condition, and it is effective. Install time as the condition, though, and the deliverable does not complete. The thirty minutes end; what you are writing does not. The problem of starting is solved and the problem of holding remains.

The completion condition also does not have to be accurate. It can be changed later. What is required is only being able to say in one line what you are making now. Say it, and the first move is settled.

The absence has an origin as well. The longer the time spent in places where conditions are supplied from outside, the thinner the experience of setting conditions yourself. Setting conditions is acquired through repetitions rather than as a skill, so thin experience is not itself a matter of disposition. Go independent while it is thin, though, and the work that cannot be started increases structurally.

And the period of delayed starting gets recorded as a period of thinking. Inside the person it is under consideration, so it does not register as doing nothing. Consideration is not counted as progress, from outside or from inside. What decreases is only the time. The mechanism of not moving as such is handled by Being unable to act is not a matter of will.

Change Comes With a Force Pushing Back Against It

What starts and stops early does so because a force is pushing back against the change. Move to the second form. Two measured biases are involved here.

The study published by William Samuelson and Richard Zeckhauser in 1988 handles status quo bias. Many decisions include an option of doing nothing, or of continuing the previous choice, and in the experiments people stayed with that option disproportionately (Samuelson & Zeckhauser, 1988, Journal of Risk and Uncertainty 1(1)). The paper shows the bias appearing at considerable magnitude in real choice data on health insurance and pension schemes among university employees as well.

The other is what Daniel Kahneman and Amos Tversky set out in 1979. People place greater psychological weight on avoiding a loss than on obtaining a gain of the same size (Kahneman & Tversky, 1979, Econometrica 47(2)).

Overlay the two and the shape of the pushing-back appears. Moving to a new form includes losing part of what the current form provides. The losing side carries a heavy coefficient and the gaining side does not carry the same weight. For a change of the same size, the internal calculation tends to tilt against the change. That is not an error of judgement but a weighting.

Prescriptions in the direction of estimating the gains more generously are plentiful. What is gained, though, lies in the future and what is lost lies in the present. The present side is concrete and the future side uncertain. Operating on the weights between the concrete and the uncertain is difficult work to execute.

Counting the losing side concretely is comparatively easy. Write out in advance what and how much you would be letting go, and it sometimes turns out smaller than assumed. Measuring the size of the object is more reliable than operating on the weights.

The mechanism by which something gets counted as a loss — where the reference point has to sit for something to become one — is handled by Where loss aversion is usable, and how the direction of the pushing-back gets changed by Mental blocks are not a defect. What is required here is one point only: the pushing-back is a normal function, not something to be removed.

The calculation has a further property: it operates independently of the content of the change. A clearly advantageous change gets pushed back too. The force operates apart from whether the content is good or bad. So from the fact of having been pushed back, nothing follows about whether the change was good. And the experience of being pushed back still gets read as evidence that stopping was right.

Not Lasting Shows Up as Three Numbers on the Business Side

Publication intervals, pricing revisions, the number of attempts at new forms.

The interval between publications. A falling frequency shows up most legibly. And the falling has a signature: it tends to fall in busy periods more than in periods of thin response. The structure of breaking when conditions disappear comes through directly.

Revisions to pricing. You decide on an increase, and the next quote goes out at the old figure. This is where the pushing-back tends to operate strongly, because raising a figure stands a loss — the possibility of losing the work — up in front of you first. And the loss side carries the heavy coefficient.

The number of attempts at a new form. How many times you tried a different way of selling, a different counterparty, a different way of making. Ending at one or two without a prior criterion is likely not a judgement but the pushing-back. The number of attempts a judgement requires is rarely one.

Is continuing with something that gets no response not waste? It can be. What divides them is whether the decision to stop came from a criterion set in advance or from a feeling that arose along the way. With a criterion, it is a judgement. Without one, it is the pushing-back with a name attached.

The criterion can only be placed before starting. Placed partway, it becomes a criterion fitted to the state you are in. Writing down, before starting, how many attempts and what results mean continuing or stopping is what separates a judgement from the pushing-back.

The periods over which they can be observed differ too. Publication intervals move by the day, pricing revisions by the job, attempts at new forms only by the month. Lay the three over the same period and whether they are moving cannot be read the same way. Starting from the publication interval, which moves on the shortest cycle, is the practical order.

Of the three, pricing revisions cannot be judged on one occasion, since whether you reverted is unobservable until the next quote arrives. So for pricing, record not the fact of the revision but the first figure issued after it. If that has reverted, the pushing-back is at work.

The three also surface more readily in the direction of deterioration. Widening intervals draw the eye, reverted figures stay in the record, abandoned attempts fix their own count. The improving side does not appear with equal force. Intervals tightening, a figure held, a third attempt entered — none of them leaves anything unless a record is kept.

And all three come with an account ready-made when they stop: the timing was bad. Busy, the market was moving, the other party was slow. The accounts are usually factual, so they cannot be refuted. Having an irrefutable account available every time is the condition for stopping at the same place indefinitely.

The Absence of a Record of Stopping Is What Manufactures the Same Place

At the centre of why not-lasting is hard to improve sits the fact that the stopped side leaves no record. What continued remains as output; what was abandoned leaves nothing behind. The article you finished remains; the publishing format you stopped after three attempts does not. The job that came in at the raised price is in the records; the decision to revert to the old figure is not a record at all.

The record is asymmetrical, so not holding how many times and at which point you stopped becomes the default state. Without it there is no material for changing the next design. You start the same form again and tend to stop at the same place.

Recording the stopping is not for reflection. It is material for deciding what to change next. Only three things are needed. How many times. What happened when it stopped. How the conditions had changed at that point.

With those three, whether the cause was pushing-back or a disappearance of conditions can be separated afterwards. If conditions changed, what needs fixing is the restoration procedure. If conditions did not change and it still stopped, it sits on the pushing-back side. The same observation — over after three attempts — yields different prescriptions.

Fix the moment of writing as well. There is not much to write immediately after stopping, because the recognition that it stopped often takes time to arrive. So write the record not at the point of stopping but at the point of trying to start again. The work of recalling what happened last time becomes the material itself.

Fix the location too. Scattered records do not get reread when you start again. Put it where you first open when starting the same form again. For publishing, where the drafts live; for pricing, beside the quote template. Match the place of the record to the place where the decision happens.

There are cases where no record is needed: where it ended as a judgement, the fit having been settled. With the judgement complete you will not start that form again, so the record has no use. Records earn their place only where the same form might be started again.

Little needs to be written. Three lines will do. Long records stop getting written, and a record about not-lasting that itself does not last is not a rare accident in this territory.

Records have a side effect too. A list of abandonments becomes material for the pushing-back in its own right. Starting a fourth attempt, the records of the previous three are in view, and that material works on the side of making the next start heavier. So write into the record not only the point of stopping but what became clear at that point. With what became clear laid out, the same record gets read on the side of the next design.

Length Does Not Tell You Whether It Lasted

Whether the direction of the pushing-back has moved shows in three places.

Days from breaking off to resuming. Not the days sustained but the interval between breaking and starting again. A shortening figure reads as establishment progressing. Breaking off did not much affect the trajectory in the measurement either. What to measure is not consecutive days but the speed of return.

Whether a decision enters immediately before execution. Deciding each time whether to do it means automaticity is still low. Beginning without deciding means the binding to context is in place. This cannot be measured in days, and recalling the moments before execution that day settles it.

Whether a record of abandoned attempts exists. For things that did not last, is there a record of how many attempts, what happened and how it stopped? With one, the next design can change. Without one, the next start takes the same form.

None of the three measures the quantity of will. What is measured is the speed of return, the presence of a decision, and the existence of a record. The pushing-back does not disappear, so whether you beat it is not the object of the judgement. Not victory or defeat, but only whether the direction of the force has moved.

Stopped in the third week, back in a fortnight, over after two attempts. What sits there is, in every case, a record of how long something lasted. Length specifies nothing about what to change next. From the same third week, one route leads to restoring conditions and another to setting a completion condition. What decides between them is not the number of days but what had disappeared at the point of stopping.

The systematising prescriptions work correctly on their own terms. Lowering the barrier acts on the starting side, and keeping a record of the stopping acts on the design that follows it. The two land at different points in the sequence, so both can be held, and either alone still moves things. With only one, the entrance for the next attempt returns to the same position as last time.

Stopping itself is normal in this structure. The pushing-back does not disappear, conditions keep moving, and completion conditions have to be set by you. What can be changed here is whether the fact of stopping becomes a record. How many attempts, what happened, how the conditions had changed. In a form where those three lines remain, the fourth attempt starts differently from the third.

The pushing-back and the changing of its direction are handled from the side of the order of cognition by Reading unable to change through cognitive science, and the order of changing itself by Self-transformation fails because the order is reversed. The problem of the estimate before starting is carried by Starting while lacking confidence, and the mechanism by which the decision to publish stalls by Being preoccupied with other people’s eyes comes from leaving the verdict on their side.