Home › Patterns › Rewriting Messages
Why does it take me half an hour to send a three-line message?
The short answer
Because the message is the only part of the exchange you can still change, and text will let you keep changing it forever. Rewriting is what care about words looks like in a medium that carries no voice. The useful question is which of those edits are still changing the message, and which stopped changing anything twenty minutes ago.
Four sentences people say about their own messages, usually to nobody.
- "It is four lines. I have been doing this since twenty past nine."
- "I deleted that sentence and typed it back in exactly the same."
- "I read it out loud to check whether it sounded cold."
- "I sent it, and then I read it again like somebody else had written it."
How the seminar finds what the extra twenty minutes are actually doing
The message needs to say three things. That you cannot make Saturday, that you are sorry, and that you would like to find another time. You know all three. You have known all three since the invitation arrived on Tuesday.
At twenty past nine you type it. At twenty-two minutes past you delete the opening word, because "Hey" is too casual for this and "Hi" is too flat and nothing at all reads as cold. At half past you have a version you would describe as fine. At twenty-five to ten you read it aloud in an empty kitchen, listening for something you cannot quite name. At twenty to you screenshot it and send it to a friend, who replies within a minute that it is fine, and you agree, and you do not send it. At five to ten you delete the second sentence and type it back in with the same words in the same order. You send it at three minutes past ten, and you read it again at four minutes past, and it looks different now that it is gone.
Thirty-five minutes. Four lines. Nobody watched any of it happen and you would find it difficult to explain to the person who received it, who read it in nine seconds while walking somewhere.
The word people reach for here is overthinking, and it is close to useless. It names the volume of thought and says nothing about the shape of it, which is why nobody has ever changed anything by being told they overthink. It also carries a small insult inside it, and the insult lands on something that is often a genuine strength. Writing well is a real skill, and one of the reliable markers of somebody who has it is that they revise.
So this page will do something narrower and more useful than telling you to hit send, which is to separate the part of that half hour that was writing from the part that was something else wearing writing's clothes, and then to look at what the something else is actually doing. Those two parts have a boundary, the boundary is findable, and most people have never once looked for it.
What Do Thirty-Five Minutes on Four Lines Actually Look Like?
Like working, from the inside, right up until you notice you are on version nine and version nine says what version two said.
The opening is usually real work. There is a thing to communicate, it has a shape, and getting it into four lines takes a couple of goes. Nobody would call that a problem. The interesting part starts a few minutes later, when the message is basically correct and the editing carries on anyway.
From there it follows a small number of moves, in more or less this order. You change a word to a synonym and change it back. You take an exclamation mark out because it looks eager, then put it in because the sentence without it looks curt. You add a softener at the front, then a second one, then cut the second because two softeners in one message reads as apologizing. You reorder two sentences to control which one they see first. You read the whole thing aloud. You put it in front of a friend. You reread the thread above it to check what register you have been using with this person for the last four months. Then, having established that the message is fine, you sit with it for another six minutes without touching it.
These are the sentences people actually run during those thirty-five minutes, in the phrasing they use when nobody is listening.
- "That sounds passive aggressive. Does that sound passive aggressive?"
- "If I put a period there it looks like I am annoyed."
- "This is the fourth time I have typed this exact sentence."
- "I will just leave it and look at it again in a minute."
- "Why is this so hard, it is a text message."
- "If they take this the wrong way I am not going to be able to fix it."
Read them slowly. The one to notice is whichever produces a small physical jolt because it is yours word for word.
Three details travel with this and almost never get said out loud. The first is that you can usually name your register with each person precisely, which is real knowledge about a real human being gathered over months. The second is that the final version is very often close to the first version, a fact people discover with some embarrassment the first time they keep their drafts. The third is the arithmetic that runs around ten at night, where you work out how long you have spent and what you would think of somebody who told you they had done that. Most people stop the arithmetic before finishing it.
The private shame sits on the third detail. It almost never sits on the first two, which are the two that make you good at this.
Why Is the Message the Only Thing You Can Still Change?
Because everything else in the exchange has already happened or has not happened yet, and neither is available to you.
You cannot change how the last conversation went. You cannot change what mood they will be in when the message arrives, whether they are on a train, whether they have just had a bad call, or how they read punctuation. You cannot change what they already think of you. All of that sits outside the window entirely.
Inside the window there is a text field with a cursor in it, and that field will accept an unlimited number of revisions at no visible cost. It is the only surface in the whole situation that responds to effort. Effort that has nowhere else to go will go there, because that is what effort does.
This is worth sitting with before any of the roots below, because it explains the strange proportions. The half hour tracks something other than the importance of the message. What it tracks is how much you care about the outcome multiplied by how little of that outcome you can touch. A message about a parking permit takes eleven seconds. A message to somebody whose reading of you feels unsettled takes half an hour, and it takes half an hour because the wording is the only variable left in an equation you would very much like to solve.
What Do Researchers Call This, and Where Do the Names Stop?
There is no single tidy label for it, and five research vocabularies each describe a real piece with precision.
Editability and reviewability name the affordance underneath everything else. Clark and Brennan set out in 1991 a list of the properties a communication medium does or does not provide, and two of them matter enormously here. Revisability is whether a sender can change a message privately before anybody receives it. Reviewability is whether a message persists in a form that can be examined again later. A spoken sentence has neither in any usable sense: it commits you the instant it leaves your mouth and it survives only as memory, which is famously loose. A written message has both at maximum. That single pair of properties is the reason this behavior exists at all, and it is a property of the channel rather than a property of you.
Rehearsability and reprocessability name the same two capabilities from inside a framework built to compare media. Dennis, Fuller and Valacich set out media synchronicity theory in 2008, describing media by a small set of capabilities: how fast information moves, how many conversations can run at once, how many symbol sets are available, how much the medium lets a sender rehearse a message before sending it, and how much a message can be re-examined afterward. Text scores at or near the top on the last two. Their argument was about matching a medium to a task. The useful part for a reader here is that your two worst moments, the composing and the rereading afterward, correspond exactly to the two capabilities the medium is strongest on. The channel is behaving as designed. You are simply using all of it.
Impression management names what the editing is in service of, and the interesting finding is about channels rather than about personality. O'Sullivan reported in 2000 that people choose communication channels partly according to what a message would do to how they are seen, moving toward more controllable channels when the content threatens the impression they want to hold and toward richer ones when it flatters it. Read that alongside the paragraph above and the shape of your evening becomes clear. You picked the channel that gives the most control, and then you used the control.
Sender-side overestimation of transmitted meaning names the error that makes the whole thing feel so urgent, and it is older and broader than text. Keysar and Henly reported in 2002 that speakers systematically overestimated how reliably listeners would identify their intended meaning, and they did it face to face, with the listener present and everything a voice provides intact. The mechanism is simple and impossible to switch off: you cannot unknow what you meant, so your own sentence looks unambiguous to you in a way it can never look to anybody else. Strip out voice and face and the error widens. This is the ground your fear is standing on, and the fear is partly right. What happens on the receiving end of that gap has a page of its own.
The counterweight is worth stating in the same section, because the standard version of this story is too gloomy to be accurate. Derks, Fischer and Bos reviewed the evidence on emotion in computer-mediated communication in 2008 and concluded that emotional communication online is considerably less impoverished than the usual account assumes. People adapt. They compensate with the cues the channel does allow, and recipients pick up a great deal more than a bare reading of the medium would predict. The channel is thinner than speech rather than empty, which means the load you are placing on the choice between two adjectives is heavier than the medium actually requires.
Relational uncertainty names the thing that varies between people, and it is the piece most readers have never had a word for. It refers to how unsure a person is about a relationship: about their own involvement, about the other person's, and about the relationship itself. Knobloch found in work on message production that people experiencing more relational uncertainty produced messages that were less fluent, more hedged, and more effortful to write. That is a precise account of something you already know from the inside, which is that this happens with four people and with nobody else, and that the four are not the four you like most.
| What this vocabulary genuinely gives you | Where it stops |
|---|---|
| A name for the affordance, revisability, which explains why this is possible in writing and impossible in speech | It describes what the medium permits and says nothing about why you use all of it and your brother uses none |
| Rehearsability and reprocessability, two measured capabilities that map exactly onto composing and rereading | It ranks media against tasks and cannot tell you what your half hour is for |
| Evidence that channel choice tracks how exposed a self-presentation feels | It shows the motive category and cannot identify which impression you are defending |
| A well-replicated finding that senders overestimate how clearly their meaning travels | It establishes the gap is real and average, not that your specific fear about tomorrow is correct |
| Evidence that text carries more emotional information than the bare medium suggests | It corrects the gloom without lowering the stakes in the one relationship you are actually writing to |
| A measured link between uncertainty in a relationship and difficulty producing messages in it | It tells you uncertainty makes writing harder and cannot tell you where the uncertainty came from |
Here is where every one of them stops, stated plainly. All six describe a what. Two people can spend the identical thirty-five minutes on the identical four lines and be doing entirely different things underneath, and the correct next step for one is close to useless for the other. One is protecting against a written record that was used against them before. One has no idea how much tone a voice was carrying for free and is trying to do that work by hand. Nothing in editability, impression management or relational uncertainty separates those two people, and the separation is the part that has to happen before anything moves.
About this page. This describes a psychological and behavioral pattern. It is not a diagnosis and it is not medical advice. I am not a therapist, and no page can tell you which of these roots is yours. Two people who recognize themselves in everything above can still need entirely different things. If you are in acute crisis, or if what you are facing involves your safety, your body, or a condition you are already being treated for, that belongs with a licensed professional now rather than later. I would rather help you find the right thing than have you keep reading here.
When Does This Belong With Somebody Qualified Rather Than a Page Like This?
Before anything practical, because it decides whether the rest applies to you at all.
Everything here assumes an ordinary and quietly expensive situation: a person who writes well, who cares how they land, and who is spending more time on it than the messages are worth. Four situations sit outside that and are better served by a person.
If composing an ordinary message produces something closer to physical alarm than reluctance, with the body responding as though a real threat were present, that deserves attention on its own terms. If the checking has hardened into a sequence you cannot interrupt without dread, and the sequence has spread beyond messages into other small acts, that is a different shape and a different kind of help. If a flatness has settled over everything for weeks, including things unrelated to other people, this page is describing a symptom rather than the picture. And if there is any thought of harming yourself, that belongs with a licensed professional now rather than later.
One more, different in kind and worth naming carefully. If somebody in your life goes back through what you wrote and uses the wording against you later, your caution is an accurate reading of your circumstances. That is a different problem in similar clothing, and no amount of better drafting solves it.
Is Being Careful With Words Something to Fix?
No, and getting this wrong would waste your time and cost you something real.
There is a whole genre of advice that treats speed as a virtue and treats any deliberation about wording as anxiety in disguise. It does not survive contact with people who write well. Good writers revise. They revise more than other people, not less, and the revising is where most of the quality comes from. If your messages tend to land, if people quote them back to you, if you have ever been told you explain things clearly in writing, then some meaningful portion of that half hour is the reason.
There is also a real difference between deliberating and stalling, and it has been studied directly in the writing itself. Flower and colleagues published work on revision in 1986 comparing how experienced and inexperienced writers actually revise. The experienced ones detected problems at the level of what the text was doing, diagnosed them, and rewrote at that level. The less experienced ones noticed trouble at the level of individual words and sentences and made local repairs that left the underlying problem in place. Both groups were revising. Only one group was changing anything.
That distinction is the whole practical content of this page, and it is more useful than any instruction about speed. The question is not how long you spent or how many versions there were. The question is whether the later versions were doing detection and diagnosis, or whether they were sanding a surface that was already smooth.
I will say the personal version of this plainly, because it took me a long time to get out of this overthinking myself, and I see it in a lot of people. In certain circles there is a great deal of thinking and a real unwillingness to take action, and the thinking gets treated as though it were the work. It is not, and I say that as somebody who spent years there.
Which Edits Are Still Changing the Message, and Which Have Stopped?
There is a clean test, and it takes about four seconds per edit.
Ask what a reader would now understand differently. That is the whole test. An edit that changes what the reader comes away with is doing work. An edit that leaves the reader with the identical understanding is doing something else, and whatever it is doing, it is not writing.
Run it against the moves listed earlier and they sort themselves immediately.
| Edit | What changes for the reader | Verdict |
|---|---|---|
| Cutting a sentence that would have landed badly | They no longer read the thing that would have stung | Real |
| Putting the request before the explanation | They see what you want before they see why | Real |
| Naming the constraint instead of hinting at it | They stop having to guess what you actually need | Real |
| Replacing a word with a close synonym | Nothing | Stalled |
| Removing an exclamation mark, then restoring it | Nothing, twice | Stalled |
| Adding a second softener to a message already soft | They read one more sentence of hedging | Slightly worse |
| Deleting a sentence and retyping it identically | Nothing at all | Stalled |
| Reading it aloud for the third time | Nothing you did not know on the second | Stalled |
The row that does most of the work is the last one in the real column, because it is where people find they were not editing for clarity at all. Naming the constraint plainly, rather than gesturing at it, changes what the reader can do with the message. It also feels riskier than any synonym, which is exactly why the eleven drafts kept circling the soft version.
There is a second signal worth knowing, and it needs no judgment. If you have restored something you previously deleted, you have reached the end of useful editing. The message has been through the space of acceptable versions and come back. Everything after that point is oscillation, and oscillation is expensive because it feels identical from the inside to the productive part.
Why Do Two People Spend the Identical Half Hour for Completely Different Reasons?
Because what a behavior looks like from outside carries almost no information about what is producing it.
Two people can send the same four lines at the same hour after the same number of drafts, and the thing they were each protecting can be entirely unrelated. That matters practically rather than philosophically. If you devise a solution based on the cause you will get somewhere, and if you devise it based on the symptom you may not, because the same technique applied to two different causes will help one person and do nothing at all for the other.
There are lots of influencing factors that go into any one behavior, which is why deciding to be quicker rarely produces a quicker person. Something is holding the current version in place. Five candidates follow. Most people find one lands harder than the rest, and a few find two.
Does It Matter That the Message Will Still Exist Tomorrow?
For some people this is the entire root, and it is the one they are least likely to have named.
A spoken sentence is gone. It survives as an impression in somebody's memory, and memory is generous: it smooths, it forgets the exact adjective, and it usually keeps the gist and the warmth rather than the wording. A written sentence does none of that. It stays exactly as typed. It can be read again on Thursday by the same person in a much worse mood than they are in now. It can be screenshotted, forwarded, shown to somebody who was not part of the conversation, or produced in an argument in two years with the timestamp attached.
If this is your root, you will recognize a specific quality in the composing: you are not really writing to the person as they are tonight. You are writing to the worst reasonable reader of that sentence at the worst reasonable moment, and you are trying to make a message that survives that reading. That is why the drafts get flatter as they go, and why the final version is often the least alive one, because vividness is the first thing sacrificed to being unquotable.
People arrive at this from real places. A workplace where written words were collected. A relationship where something you typed came back six months later. A family where a phrase you barely remember became a permanent characterization of you. The medium genuinely does what you fear it does. What is worth examining is whether you are applying a rule made under those conditions to every message you now send, including the ones going to people who have never once reread anything you wrote.
- Group
- 8–15 people
- Length
- Five days
- Where
- Sacred Valley, Peru
- You leave with
- Action steps
- The pattern
- It is doing a job
- The misread
- A character flaw
- The job
- Protection from something
- The tell
- It repeats on schedule
Five days in Peru's Sacred Valley, built around exactly the kind of pattern you just read about.
What Are Your Words Being Asked to Carry That a Voice Would Do for Free?
This root belongs to people who are good at the spoken version and have never worked out why the written version is so much harder.
In person, a difficult sentence arrives wrapped in a large amount of free information. Your face is doing something. Your voice drops or lifts. There is a pause before the difficult part that signals it is difficult. There is a small laugh that says this is awkward and I know it. Best of all, you can see the first flicker of how it landed and repair it inside two seconds, before the misreading has time to set.
Take all of that away and every one of those jobs falls to the words. The pause becomes a line break you have to decide about. The warmth in your voice becomes whether you use their name. The signal that you know this is awkward becomes an actual clause saying so, which then has to be weighed for whether it sounds insincere. The instant repair becomes unavailable entirely, which is why the pre-send checking gets so heavy: it is the only repair mechanism the channel offers, and it has to happen in advance.
If this is your root, you are not being fussy. You are doing by hand, slowly and consciously, what a voice does automatically and for free. The relief people feel when they finally call the person is the relief of handing four jobs back to a channel that does them without thinking. And the honest counterweight from the research above applies here too: readers pick up more from text than the bare channel suggests, so the hand-work you are doing is real but you are almost certainly doing more of it than the message needs.
Why Does This Happen With Four People and Nobody Else?
Because the cost of being misread is not evenly distributed, and your behavior is tracking the cost accurately.
Make the list. Most people find that eighty percent of the lost time goes to somewhere between two and five people, and that the list is not the people they love most or the people who matter most in some general sense. It is the people where the ground feels unsettled. A new manager. A friend after something happened that neither of you named. A parent with a long history of taking things a particular way. Somebody in the first six weeks of something, where every message is still establishing what the two of you are.
The vocabulary above puts this precisely: uncertainty about where you stand makes messages harder to produce, less fluent and more hedged. What is worth noticing is the direction of causation. The wording is not the problem being solved. The unsettled ground is the problem, and the wording is the only place it can currently be worked on, so it gets worked on there for thirty-five minutes at a time.
The practical consequence is unexpected and it is the most useful thing in this section. If four people account for most of your lost hours, the useful move sits somewhere other than drafting faster. It sits in whatever would make the ground with those four people less uncertain, which is usually one direct conversation and almost never a better adjective.
Is the Editing the Only Part of This You Can Actually Decide?
This root is quieter than the others and it is the one people recognize with a wince.
You cannot make them reply. You cannot make them reply warmly. You cannot make them not be annoyed, not be busy, not have already decided something about you last week. The outcome of this exchange is genuinely outside your control, and it is going to be decided by a person you cannot reach.
What you can control, completely and privately, is the wording. So the wording absorbs everything. Thirty-five minutes of effort go into the one variable that responds, and the effort produces a real feeling of having done something, which is why it is so hard to stop. Editing feels like acting on the situation. It is the closest available substitute for acting on the situation.
If this is your root, there is a specific tell. The relief comes at the send, and it lasts under a minute. Then the same discomfort returns, because the thing you actually could not tolerate was the interval between sending and hearing back rather than the wording itself, and the editing was a way of postponing the start of that interval while feeling productive about it.
The useful move here is unglamorous. Notice that you can shorten the interval more effectively than you can perfect the message. A message that asks a clear question gets answered faster than a message that has been polished into ambiguity, and speed of reply is the thing you actually wanted.
Which Version of You Is the Message Auditioning For?
The last root is the one people find slightly embarrassing to admit, which is a fair sign it is worth reading twice.
Sometimes the message is a small performance. Not a dishonest one. The content is true and you would stand behind every line. But the drafting is doing something beyond communicating: it is producing a version of you to be read as. Warm but not needy. Relaxed but not careless. Competent without appearing to try. Funny, but only at the level of funny that does not look like effort.
You can hear this root in the specific edits. It is the cutting of a sentence that was accurate because it revealed slightly more want than you would like to show. It is the removal of an apology that was sincere because two apologies would read as weak. It is the six minutes spent deciding whether to leave in the small joke, where the entire question is what the joke says about you rather than what it does for them.
Reading a whole thread back before writing is the clearest signal of this one. You are checking continuity of character. The version of you in the last eleven messages has to be the version in this one, because an inconsistency in the persona would be visible, and being visible is the thing being managed.
None of that is vanity. Every person alive manages how they come across. It becomes expensive at a particular point, which is when the person you are performing has drifted far enough from the person writing that each message needs staging. The tell is a flat feeling after sending, even when the message was received perfectly well, because what got approved of was the staged version and you know it.
| Root | What you would notice first | What the extra drafts are protecting | What actually helps |
|---|---|---|---|
| Permanence of the record | Writing for the worst reader at the worst moment | Against a sentence being reread or produced later | Sorting who genuinely rereads from who never has |
| No tone channel | Doing four jobs a voice does for free | Against a warm thing landing as a cold thing | Adding one explicit line of tone, then stopping |
| A high-stakes relationship | Two to five names take most of the hours | Against being misread by a specific person | One direct conversation with one of those people |
| Editing as the only control | Relief at send, gone within a minute | Against sitting in the wait | Making the reply easy to give and shortening the wait |
| Rehearsal of a self | Rereading the thread for continuity of character | Against being seen at a different level than intended | Sending one message slightly more like yourself |
If none of the five lands, that is information rather than failure. The most common sixth answer is the simplest one, which is that you enjoy writing, you hold a high standard for it, and the thirty-five minutes cost you nothing you miss. If that is true, this page has been describing somebody else and you can put it down with a clear conscience.
When Did Sending Start Costing This Much?
Almost nobody can name a date, and about a third of people can name a period.
The useful question is narrower than where it came from. Ask instead whether there was ever a version of you who sent messages without this, and if so, roughly when that stopped. People who can answer usually land on one of four things. A job where written communication was formal and consequential and the habit came home with them. A relationship that ended, or nearly ended, over something typed. A stretch of being new somewhere, where every message was establishing a character that had not been established yet. Or a phone that started carrying relationships that used to happen in rooms, which changed the medium of an entire life without anybody deciding anything.
The fourth is the one people underrate. A great deal of contact that used to happen with a voice and a face now happens in a text field, which means the affordance described above has been applied to relationships that never used to run through it. Nothing about you had to change for the cost to rise. The channel changed underneath the same person.
Where you can name a period, the useful follow-up is whether the conditions still hold. A rule made in a workplace that collected written evidence is a sensible rule in that workplace. Applying it eleven years later to a message to your sister is a rule outliving its conditions, which is a different problem from a personality trait, and considerably easier to work with.
What Are You Actually Asking For When You Send It to a Friend First?
Two different requests wear the same clothes here, and telling them apart takes one question.
The first request is for a reader. You have lost your ear on this message, which happens, and a second set of eyes will catch an ambiguity you can no longer see because you know what you meant. That is a completely legitimate use of a friend and skilled writers do it constantly.
The second request is for permission. You already know the message is fine. What you want is for somebody else to hold part of the risk, so that if it lands badly the decision was not solely yours. That is also human, and it is worth being honest that it is what is happening, because it explains why the reassurance does not work.
Here is the question that separates them. When they replied that it was fine, did you send it? If you sent it, you wanted a reader. If you read their reply, felt better for a moment, and then went back into the draft, you wanted permission, and permission does not transfer. Your friend cannot know how this specific person will read this specific sentence on this specific evening, so their approval settles a question you were not really asking.
There is a cost to running the second version repeatedly. The friend becomes part of the sending process rather than a person you talk to, and it is a slow, unremarked change in what that friendship is spending its attention on. The interpretation side of this loop, where the screenshot goes out to decode a message you received, works differently and has its own page.
Why Do You Reread It the Second It Has Gone?
Because sending performs a switch you did not consent to, and it happens instantly.
Up to the moment you press send, you are the author. You know the intention, you can see the alternatives you rejected, and the message reads as one option among several. The moment it is gone, the alternatives vanish and you are looking at a finished object arriving at somebody. For the first time you are standing where the reader stands, which is why it genuinely reads as though somebody else wrote it. The perspective is new. The text is not.
The rereading is an attempt to run their interpretation before they do, and it cannot work, for the reason in the research above. You cannot unknow what you meant. Every reading you perform is contaminated by the intention, which means you can generate worry with great efficiency and certainty with none at all.
What people find is that the rereading picks up something the drafting missed, and picks it up too late to act on. That happens because of the switch rather than because the drafting failed. A new vantage point produces new observations, and the only way to have had them earlier would have been to send earlier, which is the opposite of what the thirty-five minutes were for.
The practical response is small and slightly counterintuitive. Close the thread. Not as discipline and not as avoidance, but because everything the rereading can produce from here is uncertainty without information, and the one thing that will genuinely resolve it is a reply you cannot hurry.
What Does the Half Hour Actually Cost?
Three things, and it is worth being honest that for some readers the first two are close to nothing.
The hours. Four difficult messages a week at twenty-five minutes each is roughly eighty-five hours a year. We all have the same hours in the day and there are only so many things any of us can do with them, so the question is whether this is what you want to allocate that to. For some people it plainly is. Somebody who writes for a living, or who runs most of their close relationships in text, may be getting full value. Do the arithmetic before deciding, because the number is usually larger than the estimate.
The messages that never get sent. This is the cost people underestimate by the widest margin. If composing costs thirty-five minutes, not composing becomes the cheaper option, and the decision gets made below the level of choosing. The thank-you that would have taken two lines does not go. The check-in with the friend whose father is ill waits for an evening with enough room in it. Six weeks later the silence itself has become a thing to explain, and explaining it is now a harder message than the original one would have been.
The tension. There is a particular pressure and tightness you can see in somebody who is doing a lot of thinking and not much sending, and it is visible from outside long before the person notices it themselves. Sitting with a finished message for six minutes, checking, putting the phone down, picking it up again, is a small stress response run several times a day for years. Nobody plans for that and almost nobody counts it.
And the honest counterweight. If you send fifteen messages a day easily and slow down on two a week, if the writing is good and you like doing it, if none of the roots above produced a flinch, then the arithmetic is fine and you can leave this alone.
Why Does "Just Hit Send" Make It Worse?
Because it aims at the part that is visible and the visible part is the least informative part.
It is not as easy as saying, well, why do you not just do it. There is usually more to it, it goes deeper, and there is more underneath that. The instruction assumes the only missing ingredient is willingness, when the missing ingredient is different in each of the five roots and is only something like willingness in one of them.
There is a second problem. Sending an unread message does not remove the worry. It relocates it. People who force themselves through this typically spend the twenty minutes they saved thinking about what they sent, which is the same twenty minutes with the quality control removed. It is one thing to change a behavior with force or with discipline, and another thing entirely to make the change stick, and a change that costs you accuracy without buying peace tends not to survive two weeks.
And there is a third, which is why the advice stings. Somebody who has just been told they take too long over messages now has evidence that the care they take has been noticed and read as a defect. For at least two of the roots, that is precisely the feared outcome arriving in the room. The predictable next move is more caution rather than less, and now with an additional thing to hide.
The useful work sits upstream of the instruction, at the level of which root is running, and downstream of it, at the level of one specific rule applied to one specific message tonight.
What Is a Stopping Rule That Actually Holds?
A rule has to be small enough to survive a difficult message at eleven at night, which rules out anything requiring judgment in the moment.
Here is one that works, and it comes with a warning attached that matters as much as the rule.
Pass one is for content. Does the message say the thing. Is the request in it. Is anything essential missing. Rewrite freely here, including throwing the whole thing away, because this pass is where the real work lives.
Pass two is for landing. Read it once as the recipient on a bad day. Fix anything that would genuinely sting or genuinely confuse. One fix per problem.
Pass three exists only if pass two found something. It repairs that one thing and nothing else.
Then it goes. No pass four. If you catch yourself in pass four, the rule is that you may reread but may not change any words, which is easier to keep than a rule that forbids looking.
Two safeguards make this hold. The first is the restoration signal from earlier: the moment you type back in something you previously deleted, you are done, regardless of which pass you think you are on. The second is a timer, because time is the only measure available that does not require you to assess your own judgment while your judgment is the thing under load. Five minutes for an ordinary message, fifteen for a genuinely hard one. When it goes, you send the best version you currently have.
The principle underneath is worth having, because it generalizes well beyond messages. Get the details right and limit the number of things. Three passes with distinct jobs will beat eleven passes without them, in the same way that a small number of things done deliberately beats a large number done anxiously. The difficulty here has very little to do with the rule. Sending at pass three requires a small amount of courage, every time, and there is no version of this where the courage becomes unnecessary.
The warning: do not apply this to the message that actually deserves an hour. Some messages are genuinely consequential and reward real work. You will know them because passes one and two keep finding real things. The rule exists to catch the moment when nothing is being found anymore and the passes have started running for their own sake, rather than to stop you writing carefully.
- Group
- 8–15 people
- Length
- Five days
- Where
- Sacred Valley, Peru
- You leave with
- Action steps
- The premise
- No behavior is random
- The work
- Find the cause underneath
- The change
- Built on your calendar
- The point
- Life changes through action
If reading about it is no longer enough, the seminar is where the pattern gets worked on.
What Does an Easy Version of This Look Like on an Ordinary Tuesday?
Worth describing, because most people have never pictured it and are quietly assuming the alternative is carelessness.
The easy version still involves reading messages back. It still involves cutting a sentence that would have landed wrong, still involves choosing between two openings, still involves the occasional message that gets genuine time because it deserves genuine time. Nothing about care goes away.
What changes is the shape of the tail. There is no six-minute stretch of holding a finished message. The friend gets sent things to read, not things to authorize. When it goes, the thread closes and the phone goes face down, and the twenty minutes after sending are spent on something else. And the four difficult people are still the four difficult people, but the difficulty has been moved out of the drafting and into a conversation with one of them, which is where it was always going to have to go.
The other visible change is in what gets sent at all. People who resolve this consistently report that the volume goes up, particularly the small, warm, low-stakes messages that used to lose to the cost of composing. Those turn out to have been carrying a fair amount of the relationship, and their absence was doing more damage than any badly worded sentence ever did.
How Long Before Sending Stops Being an Event?
It runs at two speeds, and confusing them is what makes people give up in week two.
Noticing is fast. The behavior happens several times a week, and once you have the pass-three rule and the restoration signal, you will catch yourself inside a few days. Catching it live is uncomfortable, because you can feel yourself entering pass six and do it anyway. That discomfort is a step rather than a failure.
Changing it is slower and runs on repetition rather than on insight. The first two weeks of sending at pass three feels careless, and it will feel careless whether or not the messages are actually worse. If you have not stretched in a long time and then go stretching, the first attempts are wonky, and being wonky is not evidence that the stretching is wrong. It rarely is the first video anybody posts that does anything, and it is rarely the first week of a new rule that feels natural. It takes time and repetition to build anything of significance, and this is a small enough change that the repetition is available several times a day.
Expect it to come back under pressure. A pattern that was doing a job returns whenever the conditions that made the job necessary return, which here means a new relationship, a new manager, or a stretch where you are unsure where you stand with somebody. That is a return of the conditions rather than a loss of the progress.
What Is Worth Trying on the Next Message You Have to Write?
One small thing, and it is closer to a measurement than a change.
Tonight, on the next message that costs you more than five minutes, keep every version. Paste each one into a note as you go. When you finally send it, put version one and the final version side by side and read them together.
Most people find the two are close. Some find them identical apart from punctuation. A few find a real and important difference, and those people learn something valuable about their own editing, which is that it is earning its time. Either result is worth having, and neither is available without the evidence.
Then run one message at three passes and stay with the twenty seconds after sending without opening the thread again. What you feel during those twenty seconds is the most diagnostic information on this page. A specific fear of it being reread later points at the first root. A sense that they will not hear the warmth points at the second. A jolt about one particular person points at the third. Restlessness and a need to do something points at the fourth. A flat, staged feeling points at the fifth.
If what landed hardest was not the practical business of sending faster, but the four names that take most of your hours, or the version of yourself the messages keep staging, our Sacred Valley seminar works at that layer. We find the actual root of the behavior first, then what is standing in the way, and then build a plan you are held accountable to afterward, because the real work only starts when you come home, to the same phone and the same four people. Everybody is screened before arriving, which keeps the group small enough that somebody who has spent fifteen years choosing their words carefully cannot spend the week doing it again. One limitation worth stating plainly: exact dates, the price and the final cap on group size are still being settled, so anybody comparing programs on those numbers this week cannot yet compare ours. Ask every program you are weighing the same questions, this one included, and decide from what comes back. Keeping tonight's drafts costs you nothing and is worth doing either way.
Questions About Rewriting a Message Before You Send It
Why does it take me half an hour to send a three-line message?
Because the message is the only part of the exchange you can still change, and text will let you change it indefinitely. Speech commits you the moment you open your mouth. Writing does not commit you at all until you press send, so every worry about how you will be read has somewhere to go, and it goes into the wording. The half hour is that worry finding the only door available to it.
Is rewriting a message before sending it a bad habit?
Not on its own, and anybody telling you otherwise has skipped the part that matters. Careful writers are careful because they revise. The people whose messages land well are almost always people who read them back first. What turns a skill into a cost is the point where the edits stop changing the message and start restaging the same content, which is a specific, findable moment rather than a matter of temperament.
How do I know when an edit is still improving the message?
Ask what a reader would now understand differently. A real edit changes the meaning, the request, the order of information, or removes something that would genuinely land wrong. A stalled edit swaps a word for a synonym, moves a comma, softens a sentence that was already soft, or restores something you deleted two drafts ago. If you cannot say what the reader gains, the edit is doing something for you rather than for them.
Why do I read my messages out loud before sending them?
Because you are trying to hear the tone the reader will supply, and reading aloud is the closest available simulation of a channel the message does not have. It is a genuinely useful check for two or three passes, since your ear catches abruptness your eye misses. Past that it stops giving you new information, because you already know what you meant and cannot unhear it.
Why do I screenshot a message and send it to a friend before sending it?
Usually to borrow a reader, which is reasonable, and often to borrow permission, which is the part worth noticing. A friend can tell you whether a sentence is unclear, and that is real help. What a friend cannot do is guarantee how the actual recipient will take it. If their verdict settles you for four minutes and then you reopen the draft, you were asking for reassurance rather than for editing.
Why do I reread a message right after I send it?
Because sending removes your control over the words and immediately hands you the reader's position instead. For the first time you are looking at it as text arriving rather than as text being made, so it reads like something a stranger wrote. The rereading is an attempt to run the recipient's interpretation before they do, and it cannot succeed, because you already know what you intended.
Why does this only happen with certain people?
Because the cost of being misread is not the same with everybody. A message to a courier has no stake in it. A message to someone whose opinion of you is unsettled has a great deal. Research on relational uncertainty found that people less sure where they stand in a relationship produce messages that are less fluent and more hedged. The pattern is following the stakes rather than following your personality.
Is this the same as overthinking?
Overthinking is the general word and it explains almost nothing, which is why it never helps anybody. This is a specific behavior with a specific medium, a specific trigger, a stopping point you can locate and a cost you can count. Calling it overthinking puts it back in the box marked personality, where it has probably been sitting for a decade without changing.
What is editability, and why does it matter here?
It is the property of a medium that lets a sender revise a message privately before anyone receives it. Communication researchers list it alongside reviewability, which is the property that lets a message be re-examined after the fact. Speech has neither in any usable form. Text has both at maximum. Almost everything on this page follows from a channel that permits infinite private revision and permanent later inspection.
What is rehearsability?
A term from media synchronicity theory, set out by Dennis, Fuller and Valacich in 2008, for how much a medium lets a sender fine-tune a message before sending it. Text scores at the top. The same framework names reprocessability, which is how much a message can be re-examined afterward. Text scores at the top there too. Your two hardest moments correspond exactly to those two capabilities.
What is reprocessability?
The extent to which a message can be re-examined after it is delivered. A spoken sentence is gone once it is said, recoverable only through memory, which is forgiving. A written sentence stays exactly as written and can be read again next week by someone in a completely different mood. That property, more than any personal fragility, explains the particular weight a written sentence seems to carry.
Do people really misjudge how their tone will land?
Consistently, and not only in writing. Keysar and Henly reported in 2002 that speakers overestimated how reliably listeners would grasp their intended meaning, even face to face with the listener present. The reason is that you cannot unknow what you meant, so your own words look unambiguous to you. Removing voice and face from the channel widens an error that was already there.
Is text as emotionally thin as people say?
Less than the standard account suggests. Derks, Fischer and Bos reviewed the evidence in 2008 and concluded that emotional communication online is not as impoverished as commonly assumed, since people adapt and compensate with the cues available. That matters for you in a specific way. It means the channel is thinner rather than empty, and the crushing weight you place on individual word choice is heavier than the medium actually requires.
What is impression management, and does it apply to a text message?
It is the ordinary process of shaping how you come across, and it applies to every message anybody has ever sent. The relevant finding is about channel choice. O'Sullivan reported in 2000 that people select channels partly by what a message would do to how they are seen, favoring more controllable channels when the content threatens the impression they want to keep. Writing is the most controllable channel there is.
Why is a written message harder to send than saying the same thing out loud?
Because saying it out loud comes with a voice, a face, a pause and an immediate chance to repair. Every one of those does work for free that your words otherwise have to do alone. In writing, a single word choice carries what a change of tone would have handled invisibly. The task did not get easier when the channel got thinner, and your instinct that the words matter more here is correct.
Does the fact that the message can be reread later change how I write it?
It changes it enormously, and most people have never named this to themselves. A sent message is a durable object. It can be read again next month by a person in a worse mood, screenshotted, forwarded, or produced in an argument two years from now. Some of that fear is disproportionate. The underlying observation is accurate, and it is a property of the medium rather than a distortion of yours.
What is relational uncertainty and what does it have to do with writing?
It is the degree to which a person is unsure about a relationship: about themselves in it, about the other person, and about the relationship as a whole. Knobloch found that higher relational uncertainty was associated with messages that were less fluent and more hedged, and that took more effort to produce. It gives a precise account of why four people in your life cost you twenty minutes each and everybody else costs you nothing.
Why do I delete and retype the exact same sentence?
Because the sentence was never the problem. Deleting produces immediate relief, since the risky thing is gone. Retyping produces the same sentence, since it was the accurate one. The loop can run six times without anybody noticing, and it is the clearest available signal that the editing has stopped being about the message. Nothing on the screen is changing anymore.
Should I just hit send without editing?
No, and that advice fails for a reason worth understanding. It treats the editing as the problem when the editing is often the good part, and it asks you to give up a real skill in order to solve a discomfort. It also tends to backfire: people who force themselves to send unread messages usually spend the recovered twenty minutes worrying about what they sent. Nothing was gained and the accuracy was lost.
What is a stopping rule for editing a message?
Two passes with a purpose, then send. The first pass checks whether the message says the thing. The second pass checks whether anything in it would land wrong on a bad day. A third pass gets one job only, which is fixing something the second pass genuinely found. Any pass after that is not permitted to change words, because by then nothing you change is changing the message.
How many drafts is normal?
For an ordinary message, one or two. For something difficult, three or four, and any careful writer would recognize that. The number itself is less informative than what the later drafts do. Four drafts that each shift the meaning are four useful drafts. Eleven drafts where drafts five through eleven cycle between two acceptable phrasings are six drafts and five rounds of something else.
Why does the message I finally send often look like the first one?
Because the first version was usually accurate, and accuracy is what you were unconsciously testing against every time you rewrote it. This is the single most common report from people who start keeping their drafts. It is worth taking seriously rather than as an embarrassment. It tells you your instinct for wording is sound and that the extra time was going somewhere other than the wording.
Is this the same as reading into messages I receive?
They are relatives and they run in opposite directions. Reading into a received message is interpretation work on somebody else's words, and it has its own page. This is production work on your own words, before anybody has seen them. A person can do one and not the other. When both run in the same relationship, the exchange becomes very expensive for one person and invisible to the other.
Is this the same as rehearsing a conversation in my head?
No, and the difference is the medium. Rehearsing runs a spoken exchange forward and backward in your head, where the other person also has lines. That behavior has its own page. This one has no other person in it at all. You are working on an artifact you control completely, which is why it can absorb thirty-five minutes without ever reaching a natural end.
Is this the same as avoiding phone calls?
They come from the same property of writing and they show up differently. Preferring text is a channel choice, and the pile of unreturned calls that follows has its own page. This page is about what happens once you are already in the channel you prefer. The relief of choosing writing and the cost of composing in it are two halves of the same affordance.
Does using voice notes fix this?
For some people it genuinely helps, and for others it moves the problem. A voice note removes revision, which is the whole point, and it returns tone to the channel. It also gets rerecorded eight times by exactly the people who rewrite messages eight times. If you find you record it once and send it, you have discovered something useful about which root is yours.
Why do work emails take longer than personal messages, or the other way around?
Whichever one takes longer is telling you where the stakes sit. If work costs more, the message is defending competence. If personal costs more, it is defending closeness or standing. Some people find the split runs by person rather than by category, with two colleagues and one sibling accounting for most of the lost time. That distribution is information, and it is cheap to collect.
Am I a perfectionist, or is this something else?
Perfectionism is a plausible label and it often turns out to be too broad. A perfectionist tends to be slow across many outputs. Plenty of people who rewrite messages for thirty minutes ship rough work happily in other parts of their life, and only slow down where a specific person is going to read a specific sentence. If your slowness is general, that is a different pattern with its own page.
What if my worry is justified because this person really does misread me?
Then the caution is accurate and the question changes shape. Some people genuinely do read text uncharitably, and writing carefully to them is a reasonable response to real conditions. The useful test is whether the extra time is working. If eleven drafts still get misread, the wording was never the variable, and the fix belongs at the level of channel or conversation rather than at the level of adjectives.
Does the half hour actually cost me anything?
Count it before deciding. Four difficult messages a week at twenty-five minutes each is over eighty hours a year, and there are only so many hours in the day to allocate. The larger cost tends to be the messages never sent at all, because the composing was expensive enough that not writing became the easier option. That one accumulates quietly and shows up as distance.
Where does this usually come from?
There is no single origin, which is exactly the point of this page. It can come from a house where a sentence said carelessly had consequences, from work where written words were used as evidence, from one relationship where being misread was expensive, or simply from being good with language and having a high standard for it. Those histories look identical on the screen and need different responses.
Why does being told to relax about it make it worse?
Because it aims at the visible part and the visible part is the least informative part. It is not as easy as saying, well, why do you not just do it, since there is usually more to it and it goes deeper. Being told to relax also adds a second thing to manage: the message, and now the fact that caring about the message has been noticed and judged.
How long does it take to change?
Noticing takes about a week, because the behavior happens often enough to catch quickly. Changing the stopping point takes longer and runs on repetition rather than insight. Expect the first two weeks of sending at two passes to feel careless, in the same way stretching feels wrong after a long stretch of not stretching. It takes time and repetition to build anything that lasts.
When does this belong with a professional rather than a page like this?
If sending an ordinary message produces something closer to physical alarm than reluctance, if the checking has spread into rituals you cannot interrupt without dread, if a flatness has settled over everything for weeks, or if there is any thought of harming yourself, that belongs with a licensed professional now rather than later. The same applies if somebody in your life punishes wording after the fact.
What is one thing to try with the next message?
Keep every draft. Before you send, paste version one and your final version side by side and read them together. Most people find the two are close, which tells them something they have never had evidence for. Then send at two passes and stay with the twenty seconds afterward without opening the thread again. Whatever you feel in those seconds points at which root is yours.
Want the part most retreats skip?
Join the list for the next Sacred Valley seminar: integration-first deep inner work, an intentionally small group, and tools you actually take home.
See the seminar florian@florianrolke.com