There is a sentence I hear more often than any other in a first session, and it makes something in me bristle every time.
“I’ve had years of therapy.”
It is usually said with a mixture of pride and exhaustion, the way you might describe a long illness you survived. And it is almost always followed by the line that does the real damage: change is slow and painful.
I want to take that sentence apart, because I don’t believe it. I think it is one of the most quietly destructive ideas in my profession, and I think a great many people are still sitting in rooms every week paying good money to have it confirmed.
First, the twenty-one days
You have heard that it takes twenty-one days to form a new habit. Everyone has. It is on posters in gyms and in the opening paragraph of roughly nine thousand LinkedIn posts.
It comes from a plastic surgeon.
Maxwell Maltz published Psycho-Cybernetics in 1960. He had noticed that his patients seemed to take about three weeks to stop flinching at their own reflection after surgery. A clinical impression, from one man’s waiting room, about a very specific kind of adjustment. Somewhere between 1960 and now, “it seems to take about” quietly became “it takes,” and a surgeon’s hunch was promoted to a law of nature.
When researchers finally went and looked properly — Lally and colleagues at UCL, in 2010 — they found a median of sixty-six days for a new behaviour to become automatic. But the median is the least interesting number in that study. The range was eighteen days to two hundred and fifty-four. Some people got there in under three weeks. Others were still grinding away eight months later.
Which tells you something more useful than any single figure: how long change takes depends far more on what you are trying to change than on how long you are prepared to sit in a chair.
There are two different things going on, and confusing them costs people years
Learning to drink a glass of water when you wake up is one kind of learning. It is procedural. It runs on repetition, and repetition is genuinely slow. Sixty-six days sounds about right, and there is no shortcut. Sorry.
Unlearning the flinch you feel when your partner goes quiet is an entirely different kind of learning.
That flinch is an emotional schema. It was probably laid down fast — often in a single moment, often before you were old enough to have words for it — and it has been running underneath your conscious awareness ever since, reliably, efficiently, and without ever once asking your permission.
Here is the interesting part. Emotional learning of that kind does not appear to need sixty-six days to revise. The reconsolidation research — Nader and LeDoux’s original work, Schiller’s 2010 study, and Bruce Ecker’s clinical synthesis in Unlocking the Emotional Brain — suggests that when an emotional schema is genuinely reactivated and then met with an experience that directly contradicts it, it can be rewritten. Not suppressed. Not managed. Rewritten. And in a window measured in hours.
The reason so much therapy needs endless repetition is that it never opens that window. It builds a sensible new belief alongside the old one and hopes the new one wins on volume. Sometimes it does. It takes a very long time, and the old one is still down there, waiting for a bad week.
So when people ask me whether change is fast or slow, my honest answer is: it depends entirely on which of those two things you are doing, and most people have no idea which one they are in.
Why the years happen
I should be careful here, because there is a cynical version of this argument and I don’t quite believe it.
The cynical version says: of course your therapist tells you change is slow, they are charging you by the hour. And yes, there is an economic model in which a practitioner sees someone twice a week for six years and never has an uncomfortable conversation about progress. I’d be lying if I said I’d never seen it.
But greed is not the main problem. Drift is.
Drift is what happens when there is no formulation and no goal. Nobody ever sat down and said: this is the pattern, this is where it came from, this is what we are going to do about it, and this is how we will know it has worked. Without that, the work becomes a very expensive, very kind weekly conversation. Genuinely comforting. Occasionally profound. Not, in any meaningful sense, treatment.
And here is the thing about a weekly conversation with no destination: it cannot fail. That is precisely what is wrong with it.
The part nobody says out loud
There is a further reason long therapy persists, and it is the most delicate thing in this article.
Being listened to properly, by a kind and attentive person, for an uninterrupted hour, is one of the rarest experiences available to an adult human being. Most people do not get it anywhere else. Not from their partner, not from their friends, certainly not at work. It is genuinely valuable, and I am not about to sneer at it.
But notice what happens if that hour is the only place you get it. The therapy stops being a route to something and becomes the thing itself. You are not travelling anywhere; you have moved in.
There is a second version of this that shows up constantly in couple work. One partner is in therapy. The other is not. And the one in therapy occupies a position of considerable moral advantage: I am doing the work. I am facing myself. What exactly are you doing? Attendance becomes evidence of good faith. Years of attendance becomes evidence of heroic good faith. Meanwhile nothing at home has changed, and the therapy has quietly become a way of being right rather than a way of getting better.
I want to be careful how I put this, because there is a lazy version of the point that blames the client — you’re secretly enjoying it, you don’t really want to change — and I think that version is both cruel and wrong. It is also, awkwardly, exactly the same unfalsifiable move I’ve just accused the psychoanalysts of. If you say you want to change, that’s denial. If you say you don’t, that’s proof. No.
The honest version is structural, and it is not about you. The arrangement pays out for attendance. Turning up reliably produces the sense of doing something, the relief of being heard, and the standing that comes with taking it seriously — and it produces all of that whether or not anything actually shifts. If the reward arrives regardless of the result, do not be surprised when the result takes its time.
Which is an argument for building therapy that pays out for change instead. A stated goal. A visible measure. A date at which someone says out loud: is this working, and if not, what are we doing differently? Not to hurry anyone. To make sure the hour is going somewhere.
What we actually know about going faster
This is not wishful thinking on my part. The evidence for concentrated, structured work is considerably better than the evidence for the slow version.
Compression costs you nothing. In 2018, Foa and colleagues published a trial in JAMA comparing prolonged exposure for combat-related PTSD delivered daily over two weeks against exactly the same treatment spread across eight. The compressed version matched the spaced one. Bergen’s four-day protocol for OCD points the same way.
Be careful what that does and doesn’t say. It does not say two weeks cures PTSD. It says that when you take a treatment and squeeze it into a fortnight, you lose nothing — the calendar was never doing the work. Which raises an obvious question about the sessions that come after the treatment has finished. If time itself is not the active ingredient, then “this will take years” is a claim that needs to be argued for, not simply assumed and then billed.
Sudden gains are real and they hold. Tang and DeRubeis found that a substantial minority of clients make the bulk of their improvement in a single gap between two sessions. One week. Something shifts, and it largely stays shifted. Slow, incremental accumulation is not how a great many people actually get better — it is just how we tend to describe it afterwards, because a smooth upward line makes for a tidier story than “and then, in week five, something gave.”
Long therapy is not the norm — and the exception is instructive. Across a wide range of settings, the average course of therapy runs to somewhere between three and eight sessions. The single most common number is one.
Now, the obvious objection: people leave after one session because it was no good, or because it was expensive, or because they lost their nerve. Sometimes, certainly. But when Moshe Talmon went looking at Kaiser Permanente in the 1980s, he assumed the same thing — and on following up the people he had written off as dropouts, found that a striking proportion had stopped coming because one session had done what they needed. His group’s early study put it at over half; a large Australian follow-up found around four in ten. I’d hold those figures loosely, since most of that research comes from people who champion single-session work and satisfaction is a soft measure. But it is enough to unsettle the assumption.
Either reading is awkward for the slow-and-painful thesis. If those clients got what they came for in an hour, change is not inherently slow. If they walked out because the first hour gave them no reason to return, that is not an argument for years either. And “years of therapy” turns out to describe a small and rather particular group — one selected, uncomfortably, for not having got better yet.
How we work with it
Three stages, and they run in this order for a reason.
See the pattern. Not the presenting problem — the shape underneath it. The row about the dishwasher is never about the dishwasher.
Find the trigger. Where did this shape come from, and what fires it now? This is usually the moment something goes very quiet in the room.
Change what the trigger does. Reactivate the old learning, meet it with something that genuinely contradicts it, and let it settle differently. Hypnotherapy and timeline work are good tools for this, and they are tools rather than magic — what does the work is the mismatch, not the method’s brand name.
With couples, there is a compounding effect worth naming. Two people, two sets of patterns, and each one is the other’s trigger. Shift one and the whole system moves. This is why intensive formats often outperform the weekly hour: you catch both patterns in the same room, on the same day, while they are live.
One thing I want to say directly
If you have had years of therapy and you are still stuck, please do not read any of this as an accusation.
You were not too damaged. You did not fail to try hard enough. You did not lack insight — you have probably got more insight than you know what to do with, which is rather the point. Insight is not change. You can understand your pattern in exquisite detail and still be entirely inside it.
What went wrong was the fit. You were given a slow tool for a fast problem, and then told that the slowness was the treatment working.
Change can be quick. It can even be elegant. You just need to know where to apply the leverage.
Neil Wilkie is a relationship therapist and the founder of the Duo Coaching Framework. Duo Coaching works with couples in intensive formats designed to shift patterns rather than describe them.