Storytelling That Puts the Reader in the Scene

Here’s a sentence that appears in almost every piece of marketing ever written: You’ll have the freedom to grow your business and make more money.

Read it twice and notice what happens inside you, which is nothing at all.

Nothing happens because there’s nothing in it to picture. Freedom is a word rather than a thing. Growth is a direction rather than an event. Money is an abstraction until it’s a specific amount arriving on a specific day.

Try it on somebody. Read that sentence out loud to a friend and watch their face. Nothing will happen there either, and they’ll wait politely for you to say the part that was supposed to be interesting.

Your reader can’t see any of it, and if they can’t see it, they can’t want it.

Wanting requires a picture, and a picture requires detail, and that sentence has no detail anywhere in it.

Now try a different one. You’ll stop opening your invoicing app on a Sunday evening with a knot in your stomach. That’s a smaller promise and it does far more, because the reader has been in that exact moment.

The second sentence works because it’s a scene rather than a claim. The reader isn’t being told something. They’re being shown a piece of their own week, and recognising it happens before they’ve decided anything about you.

That recognition is what every technique in marketing is chasing, and most of them never get anywhere near it.

They’re written about a customer avatar rather than about a person having an ordinary Tuesday afternoon.

An avatar is a demographic wearing a name badge. Forty-two years old, runs a small agency, wants to scale. Nobody has ever felt seen by a description like that, and nobody ever will, because it describes a category rather than a life.

That’s why so much marketing sounds interchangeable. Everybody is writing to the same invented person, built out of the same available data, and arriving at roughly the same empty sentences about growth and freedom.

Writing scenes is the way out, and it isn’t a talent.

It’s a research problem followed by a craft problem, and both halves can be learned in a week by somebody who has never written anything creative in their life.

The avatar was invented to solve a real problem, which is that writing to everybody produces writing that lands on nobody.

Somebody sensible pointed that out, and the response was to invent a single fictional customer and write to them.

Demographics got used because they were sitting there. Your analytics hand you age and location for nothing, and nobody hands you the fact that your buyer avoids opening one particular spreadsheet on a Monday morning.

The idea was sound and the execution went wrong almost immediately. People built avatars out of demographics, because demographics are what the available data contains, and demographics have almost nothing to do with how anybody feels.

Age, income, job title, and business size tell you nothing you can write from.

You can’t build a scene out of a job title. There’s no moment in it, no time of day, nothing anybody could picture themselves inside.

So the avatar sits in a document somewhere, and the writing that comes out of it is exactly as generic as the writing that came before it. The person who built it concludes that avatars don’t work.

You can prove this to yourself in about a minute.

Take your own avatar document and try to write a sentence from it that would make a real person stop scrolling. You’ll find you cannot, and the reason is that there’s nothing in there to see.

They do work, provided they’re built out of the right material. The right material is behaviour and moments, not attributes. What your buyer does on a Thursday, and what they’re avoiding, and what they’ve already given up on.

Behaviour is harder to get and it’s the only thing that produces writing anybody feels.

Attributes are easy to get and they produce writing that reads like a form. The difficulty of getting the material is exactly why so few people have it.

Consider the difference. An avatar that says small agency owner, wants more clients, gives you nothing. An avatar that says opens the pipeline spreadsheet on Monday, sees two names on it, closes it again, gives you a whole piece of writing.

The second one contains a scene.

It has a time, an object, an action, and a feeling underneath it that never had to be stated. You could write a hundred pieces from that one line and every one of them would land.

Which means the avatar work most people do isn’t wrong so much as pointed at the wrong information. You’re not trying to describe who your buyer is. You’re trying to describe what happens in their week.

You could write a hundred pieces from that one line and every one of them would land.

That change in target also changes where you go to get the material, and this is the part that trips everybody up. Demographics come from analytics. Moments come from conversation, and conversation is slower and nobody wants to do it.

That’s unwelcome news and it deserves to be said plainly rather than dressed up. There’s no tool, no prompt, and no dataset that will hand you the specific moments in your buyer’s week. Somebody has to ask them.

There’s no way around that, and pretending otherwise is why so much marketing sounds identical.

The people who write scenes are the people who talked to somebody, and everybody else is guessing from a dashboard.

The good news is that it takes fewer conversations than you’d think. Six proper conversations with real buyers will give you enough material for a year, because the same three or four moments come up again and again.

Six is the number because the fourth conversation starts sounding like the third. When you notice yourself predicting the answers, you have enough, and continuing past that point is procrastination wearing a research badge.

Those repeated moments are the ones to write about.

When four people independently describe the same Sunday evening feeling, you’ve found something that will land on a thousand people who never told you anything.

Test your own avatar right now by trying to write one scene from it. Take what you have and try to describe a specific ten minutes in that person’s week, with objects in it and a time on the clock.

If you can’t do it without inventing, your avatar is made of the wrong stuff. Inventing is where generic writing comes from, because what you invent will be what you’d feel, and you aren’t your buyer.

Here’s my current customer avatar or audience description: [paste it]. Tell me plainly whether this is built out of demographics or out of behaviour. Then list every question I’d need to ask a real buyer to rebuild it out of moments instead. The questions should be about what happens in their week, what they avoid, what they’ve given up on, and what specific tasks they dread. Give me the questions in an order I could work through in a fifteen-minute call.

Book six of those calls before you write another word of marketing. That sounds like an interruption to the work, and the calls are the work, and the writing afterward takes half as long and performs several times better.

Record them with permission and go back to the recordings rather than to your notes. Your notes will contain what you thought was important. The recordings contain the offhand sentences, and the offhand sentences are the material.

The details that make a scene work are ones you’d never have thought of, and that’s precisely why they work.

A reader accepts a detail they recognise and wouldn’t have expected, because nobody invents something that specific.

So the collection problem is really a question of getting people to say things they don’t know are interesting. Nobody volunteers the good material, because to them it’s just their life and their life doesn’t seem noteworthy.

Direct questions produce useless answers. Ask somebody what their biggest challenge is and they’ll give you a sentence they’ve said before, polished by repetition, containing no detail whatsoever. It’ll sound like a survey response, because it’s one.

The reason polished answers are useless is that polish removes exactly what you came for.

Every time somebody has told a story before, they have smoothed off the odd details, and the odd details were the whole point of asking.

What produces detail is asking about a specific occasion. Not what’s hard about invoicing, but tell me about the last time you sat down to send invoices. The first question gets a category. The second gets a Tuesday.

That’s the entire interviewing technique, and everything else is a variation on it. Take people from the general to the particular and keep them there. They’ll hand you material you could never have made up.

Notice that the second question also gives them permission to be ordinary.

A question about the biggest challenge invites a big answer, and people invent big answers. A question about last Tuesday invites the truth, because Tuesday wasn’t impressive.

Follow-up questions do the rest, and the useful ones are almost embarrassingly small. What time was it. Where were you sitting. What was on the screen. What did you do straight afterward. What did you say to yourself.

Those questions feel intrusive to ask and they aren’t. People enjoy answering them, because nobody has ever been that interested in their ordinary Tuesday, and the specificity gives them permission to say the true thing.

Silence is a tool here and most people fill it too quickly.

Ask the question, and then wait, and let the pause sit there being awkward. The thing they say to break the silence is nearly always the thing you came for.

The true thing is usually the fourth or fifth answer, not the first. The first answer is the story they tell people. Keep going, keep asking what happened next, and somewhere around the fifth exchange the real account appears.

You’ll know it when it arrives because it’ll contain something slightly odd.

They’ll mention that they made a cup of tea and left it to go cold. They’ll mention the drawer they put the letter in. That’s the detail you’re mining for.

You’ll feel a strong urge to skip past the small detail and get back to the business problem, and that urge is the enemy. The business problem is the part you already know. The cold cup of tea is the part you came for.

They’ll mention that they made a cup of tea and left it to go cold.

Write those down exactly as they said them, in their words, without improving anything. The urge to tidy up their phrasing is strong and it destroys the value, because the roughness is the part that reads as true.

They’ll also outlast your product. Details about a buyer’s week stay true for years. Everything you know about your offer expires the moment you change it, which makes this the most durable research you can do.

Ten of those details is a working supply.

You’ll use them for a year across everything you write, and they’ll do more for your conversion rate than any headline you could test in the same period.

AI is useful here in a way it isn’t elsewhere, because this is a listening job with a clear structure. You can hand it a transcript and have it pull out everything concrete, ranked by how few other people would have said it.

The same trick works on your own memory, which is useful when you can’t get a call booked. Ask yourself about the last specific time you had the problem you now solve, and refuse to answer in generalities.

You’ll find you’ve been describing your own experience as a summary for years, and the summary has no detail in it either. Your own Tuesday is as usable as anybody’s, provided you go back and look at it properly.

Here’s a transcript of a conversation with one of my buyers: [paste it]. Pull out every concrete, physical detail in it. Objects, times of day, places, actions, and the exact words they used about how something felt. Ignore anything general or summarised. Rank the details by how unlikely another person in a different situation would be to mention the same thing. I want the ones that could only have come from a real week.

Keep the ranked list somewhere you’ll see it while you draft, because you’ll forget these details exist otherwise. The whole point of collecting them is to have them in front of you at the moment you’re reaching for a generic sentence.

Add to it every time you talk to anybody. Support emails count, replies count, and a two-line comment counts if it contains one true thing. The list grows on its own once you’re looking for the right material.

With details collected, the craft problem is simple enough to teach in a paragraph.

A scene has four parts, and the parts go in a particular order, and almost nobody bothers with any of them.

The first part is a time. Not a vague one, but an actual point in a week. Sunday evening. Wednesday at eleven. The morning after a launch. Time anchors the reader, because they were somewhere at that moment too.

The second is a place with an object in it.

At the kitchen table with the laptop open. In the car outside the office. In front of a dashboard that hasn’t changed since Friday. Objects are what make a scene visible.

The third is an action, and it’s usually a small one. Opening something, closing something, refreshing something, not sending something. Small actions carry more weight than large ones, because everybody’s week is made of small actions.

Times of day carry more weight than people expect, and it’s because a time implies a life.

Sunday evening isn’t just a moment. It’s the end of a weekend and the approach of a week, and every reader knows what that feels like.

The fourth is what they told themselves. Not the feeling, which is abstract, but the sentence in their head. Maybe next month. It’s probably the subject line. Everyone else seems to have this worked out.

Put those four together and you have a scene. Sunday evening, at the kitchen table, opening the invoicing app and closing it again, thinking you’ll do it tomorrow. Anybody who has lived that reads it and stops.

The four parts also work as a checklist when a scene falls flat, and it usually falls flat for the same reason every time.

One of the four is missing, and it’s nearly always the sentence in their head.

Notice what the scene never does. It doesn’t say you’re overwhelmed. It doesn’t say you’re struggling with cash flow. It never names a feeling anywhere, and the feeling arrives in the reader more strongly than if it had been named.

Scenes are also the one place in your writing where being boring is completely safe. An ordinary Tuesday is dull to describe and startling to recognise, and the recognition is what you’re after rather than the drama.

That’s the rule that governs all of this.

Naming the emotion tells the reader what to feel, and being told what to feel produces resistance. Showing the scene lets them arrive at the emotion, and an emotion you arrived at is one you trust.

The mistake people make is putting the feeling in as insurance, because they’re worried the scene isn’t clear enough. Sunday evening, the invoicing app, and then a sentence explaining that this is exhausting. That sentence undoes the work.

Sunday evening, the invoicing app, and then a sentence explaining that this is exhausting.

That instruction is the hardest one here and it’s the one that separates writing people finish from writing people skim. Every explanation you add is a small insult, and readers feel insults without being able to name them.

So write the scene and delete the explanation. Trust the reader completely. If the scene is specific enough, they’ll do the interpreting, and the interpretation they do themselves will be stronger than anything you could have written.

Once you’ve written a few, you’ll start seeing scenes everywhere in other people’s writing, and you’ll notice how rare they are.

Almost nobody does this. That rarity is why it works so well when you do.

Length is the other thing people get wrong, and shorter is nearly always right. Three sentences is a scene. Ten sentences is a short story, and your reader didn’t come here to read a short story about themselves.

Two or three scenes across a piece is plenty. They work like anchors, dropping the reader back into their own life at intervals. The teaching in between lands because the reader is standing somewhere real while they read it.

The bareness is uncomfortable at first and it’s the correct amount of discomfort. You’re used to writing that carries its own explanation, and scenes don’t, and that means trusting somebody you can’t see to do half the work.

They’ll do it. Readers are far better at this than writers give them credit for. The evidence is in every novel anybody ever finished, none of which explained what the reader was supposed to be feeling.

Here are the concrete details I’ve collected from my buyers: [paste your ranked detail list]. Write me six scenes, each three sentences long. Every scene needs a specific time in the week, a place with an object in it, one small action, and the sentence the person says to themselves. Never name an emotion anywhere. Never explain what the scene means. Use my buyers’ own words wherever possible, and keep every scene inside an ordinary week rather than at a dramatic moment.

Read the six and pick the two that made something turn over in you. Those are the ones that’ll do the same to your reader, and your own physical reaction is a more reliable guide here than any reasoning you could apply.

Put the other four in a file rather than deleting them. A scene that does nothing today will be exactly right in three months, when you’re writing about a different part of the same problem.

Every guide to marketing writing tells you to use the second person, and the advice is right as far as it goes.

It’s also the source of a failure that nobody ever warns you about, and it costs people readers every day.

The failure is that you is only powerful when the sentence after it is true of the reader. When it isn’t, the word does the opposite of what it was supposed to do, and it does it fast.

Consider what happens when a reader meets the sentence you’re spending hours every day on social media, and they aren’t. They don’t think this is close enough. They think this isn’t for me, and they leave.

That’s the risk you take every time you use the word.

You is a claim about somebody you’ve never met, and if the claim is wrong, you’ve told them explicitly that you don’t know them.

You can watch yourself do this while reading. A sentence that describes you accurately produces a small physical reaction. A sentence that describes somebody else produces a decision, and the decision is to stop reading.

Which is why generic writing hedges, and why hedged writing feels so lifeless. Writers sense the danger and retreat into safety. You might be finding it hard to keep up. That sentence is true of everybody and lands on nobody.

Every hedge you add makes a sentence safer and weaker at the same time.

The words might, sometimes, and perhaps are all ways of avoiding the risk of being wrong, and they cost you the reward of being right.

So there’s a trade, and the trade is the whole craft of it. The more specific your you sentence, the more it lands on the people it fits and the harder it bounces off everybody else. That’s not a problem to solve. It’s a choice to make.

Make the choice deliberately, and make it toward specificity, because the people it bounces off were never going to buy anyway.

A sentence that describes forty percent of your list precisely will outperform one that vaguely describes all of it.

There’s also a way to get the specificity without the risk, and almost nobody uses it. You stop writing you and start writing about somebody else, and you let the reader recognise themselves without being addressed.

A woman I spoke to last month opens her invoicing app every Sunday evening and closes it again without sending anything. That sentence contains no claim about the reader, and a reader who does that will feel it more than if you’d accused them.

Here’s what’s happening underneath. Being described to your face is uncomfortable and it produces resistance. Overhearing a description of somebody like you produces recognition instead, and there’s no defence against recognition, because nothing is being asked of anybody.

The trick works because a reader eavesdropping on somebody else has no reason to defend themselves. They aren’t being addressed, so there’s nothing to push back against, and the recognition lands before any defence gets raised.

Being described to your face is uncomfortable and it produces resistance.

So the strongest pattern alternates. A scene about somebody else, then a sentence in the second person that only arrives after the recognition has already happened. By then the you is safe, because they’ve already agreed with it.

That sequence solves the whole problem. You get the specificity of a real person’s Tuesday, and then you get the intimacy of the direct address, and neither one is doing damage to the other.

Most pieces have one or two sentences that fail this test badly, and they’re usually near the top, where the writer was reaching hard for a hook. That’s the worst possible position for a claim that might be wrong.

A reader who’s been misdescribed in the first paragraph will never trust the rest of it, however good the rest of it is. Get the opening you sentences right and the rest of the piece has a chance to work.

Here’s a piece I’ve written: [paste it]. Find every sentence addressed directly to the reader as you. For each one, tell me how confident I can be that it’s true of most of my readers, and flag the ones that are risky assumptions. Then rewrite each risky sentence as an observation about a specific other person doing something specific, so a reader who recognises themselves can do so without being told what they’re like.

Read the rewrites and notice how much less pushy the piece becomes. It says the same things and asks nothing, and asking nothing is what lets a defended reader stay in the writing long enough to be affected by it.

Keep a few direct you sentences, particularly late in the piece, where the reader has already recognised themselves several times over. By then the direct address is welcome, because it’s confirming something rather than presuming it.

The same problem afflicts the promises you make, and this is where most marketing collapses into noise that nobody can tell apart.

You’ll grow your business. You’ll get your time back. You’ll be free of the stress at last.

None of those can be pictured, and a promise that can’t be pictured can’t be wanted. The reader agrees that it sounds nice, and agreeing that something sounds nice isn’t the state that precedes a purchase.

Read your own promises out loud and notice how little you feel while saying them. If they do nothing to you, and you built the product, they’ll do considerably less to a stranger who has never heard of you.

So a promise has to become a scene, the same as everything else.

Not you’ll get your time back, but you’ll close the laptop at six on a Friday without opening it again until Monday morning.

That’s the same promise and it does something entirely different, because the reader can see it. They can see the laptop closing. They know what Friday at six feels like, and now they know what it could feel like instead.

The construction is simple enough once you’ve seen it done. Take your abstract promise, find the specific moment in the reader’s week where the change would show up, and describe that moment instead of the abstraction.

The moment is the thing to hunt for, and it’s always smaller than you’d like.

Ask yourself what would be the very first sign, in an ordinary week, that the thing you sell had started working for them.

More money becomes an invoice for eight hundred pounds arriving on a Tuesday from somebody you’ve never met. Less stress becomes not checking your email at traffic lights. Freedom becomes taking a Wednesday off without telling anybody.

Each of those is smaller than the abstraction it replaces, and each one is worth more.

Smallness is what makes it believable, and believability is what makes it desirable, and neither of those things is available to a big vague promise.

Small promises also survive contact with a sceptical reader. A person who has been disappointed before can imagine a quiet Friday. They can’t imagine their whole business changing, because nothing has ever changed their whole business.

There’s a discipline here that people find hard, which is resisting the temptation to promise the transformation.

Somebody’s whole life changing is a bigger promise than a quiet Friday evening, and it converts worse, and it converts worse for a reason.

The reader doesn’t believe their whole life will change, because it never has before. They do believe a Friday evening could be different, because that’s a small enough claim to be plausible, and plausible is the only thing that produces action.

So the working rule is to promise the smallest specific thing they’d care about. Not the whole outcome, but the first visible sign of it, described precisely enough that they can see themselves inside it.

The reader doesn’t believe their whole life will change, because it never has before.

Big promises also attract the wrong buyers, which nobody warns you about. The people who respond to a promise of total transformation are the people who have bought that promise many times before, and they’ll refund you as well.

That promise also has an honesty advantage, which is worth having. You can deliver a quiet Friday evening. You can’t deliver a transformed life, and a buyer who was promised one and received the other will ask for their money back.

The small promise, kept, produces a customer who trusts you and buys again.

The large promise, half-kept, produces a refund and somebody who tells other people about it. The arithmetic on that isn’t complicated.

Those four words are so common in marketing that readers have stopped processing them entirely. They pass over them the way you pass over a legal disclaimer, and the sentence containing them does no work at all.

Which means they’re free to remove. Take them out of a piece and put a scene in their place. The piece gets shorter, the promise gets stronger, and nothing at all has been lost.

Here are the promises I make in my marketing: [paste them]. My buyers are [describe them]. For each promise, find the specific moment in a normal week where a buyer would first notice that promise coming true. Describe that moment as a scene with a time, a place, an object, and a small action. Keep every scene modest and believable rather than dramatic. Then tell me which promises I should stop making entirely because nobody could picture them.

That last question usually removes half your promises and improves the other half. The ones that go were doing nothing anyway, and their absence makes the remaining ones easier to see and harder to ignore.

Watch what happens to your refund rate afterward, because it drops. Buyers who bought a small specific promise get what they were promised, and people who get what they were promised very rarely ask for their money back.

None of this has to be guesswork, which is the useful thing about it.

Whether a reader saw themselves in your writing is a testable question, and the test is simpler than any analytics you could set up.

Ask them. Send a piece to twenty people and ask one question in reply. Which part of this sounded like your week? Six people will answer, and their six answers will teach you more than a month of split testing.

That question works because it’s easy to answer and hard to answer politely.

A person can’t be vaguely kind about which sentence sounded like their week. They either point at one or they tell you that none of them did.

The answers cluster, and the clustering is the finding. When four of them point at the same sentence, that sentence is doing the work, and everything else in the piece is scaffolding around it.

Twenty people is enough and you don’t need a statistician to interpret it. You aren’t measuring a rate. You’re finding out which sentence a real person recognised, and six answers will tell you that clearly.

Then write the next piece around that sentence, and ask again.

Within a few rounds you’ll have a small set of scenes that reliably land, and those scenes will carry your marketing for years.

Do it on a piece you’ve already sent rather than waiting for the next one. The replies will arrive within a day, and you’ll know more about your readers by Friday than you learned in the whole of last year.

There’s a second signal that costs you nothing to collect, which is what people quote back to you.

When somebody replies to an email and repeats one of your lines, that line got inside them, and you should note it down.

The lines people quote back are almost never the ones you were proudest of. They’re the small concrete ones, the Tuesday details, and their being remembered is proof that the scene did what a scene is supposed to do.

Watching for that’s a habit worth building, and it takes no effort at all.

Every reply you get contains a small piece of evidence about which part of your writing was visible, and most people delete the replies without reading them properly.

Keep a running note of the quoted lines and read it every few months. You’ll see the same three or four details coming back, and those are the ones to build your next launch around.

There’s a harder check you can run on yourself, and it’s the one that catches generic writing before it goes out. Read your piece and ask whether a competitor could have written the same sentences.

Keep a running note of the quoted lines and read it every few months.

If they could, the writing is generic, however good it sounds. A sentence that could appear in anybody’s marketing is doing nothing for you, and the specificity that would make it yours is sitting in your interview notes.

That check works because generic writing is always the result of guessing rather than collecting. Where you guessed, you produced what everybody else produces, because everybody who guesses guesses in roughly the same direction.

So the presence of a stolen-sounding sentence is a signal about your research rather than your writing.

Somewhere behind that sentence is a detail you didn’t have, and the answer is a conversation rather than a rewrite.

So the loop is short and you can run it forever. Collect real moments, build scenes, ask which ones landed, then collect more moments about the ones that did. Everything else people do to improve their marketing sits downstream of this.

People will tell you they liked it because that’s a polite thing to say, and you’ll learn nothing. Asking which part sounded like their week forces them to point at a sentence, and a pointed-at sentence is data.

Most readers enjoy being asked, because nobody asks them anything. The question itself does a little work for you, since it tells them somebody is trying to describe their life accurately rather than sell them something.

Here’s a piece I’ve written: [paste it]. Go through it sentence by sentence and flag every one that a competitor in my niche could have written word for word. For each flagged sentence, tell me what specific detail from a real buyer’s week would make it unmistakably mine. Then tell me which questions I’d need to ask a customer to get those details, phrased so they’d produce a scene rather than a summary.

Expect a lot of flags on the first run, and don’t take it hard. Every writer’s first draft is full of sentences that belong to nobody, because generic phrasing is what arrives when you’re thinking about the argument instead of the person.

The second draft is where the person goes back in. That’s what all of this comes down to in the end, which is putting a real Tuesday back into writing that had drifted into describing a category.

All of this comes down to one thing being true, and it’s true, and it’s easy to forget while you’re staring at a blank page.

Your reader is a person having an ordinary week, and their week is full of specific moments you could describe.

They aren’t a market segment, and they have never once in their life thought of themselves as one. They’re somebody who closed a laptop this evening with something unfinished on it, and who will think about it again at eleven tonight.

Everything in your marketing that has ever worked, worked because a person recognised their own life in it. Every technique anybody has ever taught you is a different route to that same moment of recognition.

Write that, and they’ll read anything you put next to it.

Write the segment, and you’ll produce sentences that four hundred other people in your niche have already written, and your reader will pass over all of them without stopping.

The difference between those two outcomes isn’t talent and it isn’t cleverness with words. It’s whether you went and found out what happens to somebody on a Tuesday, or whether you sat at your desk and guessed.

Guessing feels productive because it produces sentences quickly, and speed feels like progress when there’s a deadline.

The sentences it produces are the ones your reader has already scrolled past a hundred times this year.

Six conversations is what it costs. Ten details is what you get. The writing that comes out afterward describes somebody’s real life back to them. That’s the only thing anybody has ever wanted from a piece of writing.

This lesson is video’s native language. A scene with a time, an object and an action is exactly what an AI video prompt needs.

The details you collected from buyers translate straight into Wan prompts: Monday morning shown through the light, the spreadsheet on the screen, the two names, the laptop lid closing. Describe what a camera would see, never what the viewer should feel.

Then resist explaining it. A caption that names the moment, such as “Monday, 8.15am”, lets viewers recognise themselves. A caption that tells them how to feel undoes the scene.

Here are concrete details I’ve collected from my buyers: [paste them]. Turn the three strongest into AI video scenes. For each, write a Wan 3.0 prompt with one subject, one setting, one simple action, the time of day shown through the light, a fixed or slow camera, and no readable text. Then write one caption line that names the moment without explaining it.