Video Marketing · Storytelling That Sells · Lesson 5 of 10

Storytelling With Proof Built In

Numbers bounce off people. Put the proof inside the story, where doubt actually happens.

7 parts7 copy-and-paste promptsVideo scripts included29 min read
Storytelling With Proof Built In cover
Storytelling That Sells · Lesson 5

Why This Matters

Watch somebody read a sales page and you’ll see something that should worry you.

They read the opening. They slow down over the part about their problem. Then they hit the section with the numbers in it, and their eyes drop straight to the bottom of it.

They didn’t decide to skip your evidence. Nothing that deliberate happened. Their eyes just slid off it, the way eyes slide off a legal footer, and the strongest material you had went past them without registering.

That’s a painful thing to learn, because the proof was the part you were proudest of.

You have the results. You have the screenshots. You went and got the testimonials, and the reader skimmed all of it in under two seconds.

The reason has nothing to do with whether your proof is good. It has to do with where you put it. A number sitting on its own has no context, and a reader with no context has no way to judge whether the number means anything.

Think about how you read a claim yourself.

Somebody tells you their client tripled sales, and your first thought isn’t admiration. It’s a question. Tripled from what? Over how long? Was the client already halfway there?

Those questions arrive automatically and they arrive before belief does. If your writing doesn’t answer them, the reader parks the number in the same place they park every other number they’ve been shown, which is nowhere.

Proof only works when it’s attached to a person the reader can picture.

A number inside a story gets believed because the story supplies the answers to all those questions before the reader has to ask them.

There’s a second thing going on that makes this worse. Proof sections tend to look different from the rest of the page. Different font, boxes, logos, star ratings. The reader has learned that anything formatted that way is an advertisement, so they skip the formatting before they read a word of it.

So the whole problem is one of placement and packaging rather than of collecting more evidence.

You almost certainly have enough proof already. What you don’t have is proof the reader can absorb, and those are two different things.

Part 1 of 7

The Reason Numbers Bounce Off People

01

A number is an abstraction, and abstractions are hard for people to feel anything about.

Tell somebody a figure went up forty percent and they’ll nod politely. They have nowhere to put the forty percent, and no reason to care where it came from.

That’s not stupidity or laziness on their part. It’s how everyone reads. A percentage has no weight until it’s attached to something with a face on it. Your reader is being sensible when they refuse to spend energy on a claim with nothing anchoring it.

You can prove this to yourself in ten seconds.

Try to remember a single statistic from any sales page you’ve read this year. You won’t manage it. Now try to remember a story from one, and you probably can.

There’s a second problem with bare numbers, and it’s worse than the first. Your reader assumes you picked the best one. They know you have a spreadsheet somewhere with a hundred results in it, and they know which one you chose to show them.

Think about the last time somebody showed you a stack of five-star reviews.

You didn’t become more convinced. You started looking for the one-star ones, because a wall of praise triggers the part of you that assumes something is being managed.

That suspicion is well earned. Everybody does select their best number, and readers have watched it happen for years. A big figure standing alone therefore reads as advertising, and advertising gets discounted before it gets read.

You can watch this happen to yourself.

Somebody sends you a screenshot of a bank balance and you feel nothing. Somebody tells you their mother stopped asking when they were going to get a real job, and you feel something immediately. Same result, entirely different effect.

Here’s the strange part. A smaller number is often more believable than a bigger one, precisely because nobody would invent it. Somebody who made an extra three hundred dollars in a month sounds like a real person having a real month.

Somebody who made forty thousand sounds like a headline.

The reader doesn’t disbelieve it exactly. They file it under things that happen to other people, and once it’s filed there it can’t do any work on their decision.

So the first correction is often to lead with a modest result rather than your best one. That will feel like leaving your strongest card on the table, and it isn’t, because your strongest card was never being played in the first place.

So the first correction is often to lead with a modest result rather than your best one.

None of this means you should hide your best results forever. It means you have to earn the right to show them. A large number shown late, after a reader already believes you, reads as the upper end of what happened. Shown early, it reads as bait.

The third problem with bare numbers is that they raise questions instead of settling them. A figure with no context makes the reader wonder. A story with a figure inside it makes the reader nod, because they can see how the figure came about.

Wondering is the enemy here.

A wondering reader stops reading forward and starts reading backward. They check your earlier claims against this new one, and that’s how a person talks themselves out of buying.

What you want instead is for the number to arrive after the reader already believes it. That sounds backwards and it’s the whole trick. Tell the story first, and the number becomes confirmation of something they’ve already accepted.

💡
Tip

Take the biggest number in your marketing and ask whether a reader could picture the person it belongs to. If they can’t, the number is doing nothing for you. Attach it to a name, a starting point, and a week, and it starts working.

There’s a simple test for whether a number is landing. Read the sentence before it and the sentence after it, and ask whether either one gives the reader any reason to trust the figure sitting between them.

Read the same three sentences the way a stranger would, which means assuming you have something to gain by lying. That isn’t cynicism, it’s simply the position every reader occupies, and pretending otherwise is what produces marketing nobody believes.

Most of the time, neither does.

The number is announced, and then the writing carries on to the next point as though the reader has accepted it. They haven’t. They’ve noticed that you moved on quickly, which tends to confirm their suspicion.

Copy-and-paste AI prompt
Here are the results and numbers I use in my marketing: [paste them all, with any context you have]. For each one, tell me what questions a sceptical reader would ask before they’d believe it. Then tell me what information I’d need to add so those questions get answered before they’re asked. Last, tell me which of these numbers I should stop using entirely because they’re too large to be believed by somebody starting from nothing.

Expect that last answer to sting a little. The number you’re most proud of is usually the one doing the least work. It sits furthest from your reader’s own situation, which makes it the easiest one for them to dismiss.

Keep it for later in the piece, once belief has been established by something smaller. A big number at the end of a page reads as an upper limit. The same number at the top reads as a claim you’re trying to get away with.

✅ Do this
  1. Tell somebody a figure went up forty percent and they’ll nod politely.
  2. Try to remember a single statistic from any sales page you’ve read this year.
  3. Tell the story first, and the number becomes confirmation of something they’ve already accepted.
  4. Read the same three sentences the way a stranger would, which means assuming you have something to gain by lying.
Part 2 of 7

Proof That Lives Inside a Story

02

The alternative to announcing your proof is burying it inside something the reader wants to read.

Not hiding it, which would be pointless, but placing it where it arrives as part of an account of something that happened.

Here’s the difference in practice. The announced version says a customer increased her open rate by twenty percent. The buried version tells you about a woman who stopped sending Tuesday broadcasts because nobody opened them, and what she tried instead.

By the time the twenty percent shows up in the second version, the reader has met her.

They know she was frustrated. They know what she changed. The number is no longer a claim, it’s the ending of a story they were already following.

That change in status is what makes it believable. A claim invites resistance because it asks you to accept something. An ending doesn’t ask anything, because you were reading toward it and you arrived there yourself.

Notice what the story version never does.

It never says the product is good, and it never asks the reader to believe anything. It just recounts what happened, and the reader draws the conclusion on their own, which is the only conclusion they’ll ever fully trust.

A conclusion a reader reaches themselves is one they’ll defend. A conclusion you handed them is one they’ll argue with. That difference is why story-shaped proof survives scepticism that a bare claim never could.

Building proof this way requires you to collect different information than you currently collect.

Most sellers ask their customers for a testimonial and get back a compliment, which is useless. You need the story instead.

Compliments are what you get because compliments are easy to give. Somebody who liked your product will happily say so in eight words. Those eight words will do nothing for a stranger trying to decide whether to trust you.

A compliment tells the reader that somebody liked you.

That was never in doubt and it changes nothing. The story tells the reader what somebody’s situation was, what they did, and what changed, and that’s a thing another person can copy.

The reason you get compliments is that you asked for them. Asking somebody what they thought of your product invites an opinion, and an opinion about you is the least useful thing a customer can give you.

The reason you get compliments is that you asked for them.

So the questions you ask your customers have to change. Stop asking what they thought of the product. Ask where they were before they bought it, what they’d already tried, what the first week was like, and what’s different now.

Those four questions produce material you can use. The first one gives you a starting point the reader can match themselves against. The second proves you’re not selling them something they already failed at. The third tells them what to expect.

The fourth is where your number lives, and by then the reader has enough around it to judge it fairly.

The figure arrives at the end of a sequence they’ve been following, and it lands as the natural result of everything above it.

You’ll get pushback on this from customers who want to be brief. Somebody will write back saying the product was great and leave it there. Reply and ask the one follow-up question, and most of them will answer it, because people like talking about themselves.

There’s an extra benefit to collecting stories instead of compliments, and it shows up in your writing rather than in your marketing.

Stories give you specifics you’d never have invented, and specifics are the thing your writing is always short of.

🧠
Idea

Send one email to your last twenty buyers asking a single question: what were you doing about this problem before you bought? Nothing else. The answers are the opening lines of twenty stories, and they cost you one email to get.

One question gets answered far more often than four questions do. People see a list and put it off until the weekend, and the weekend never arrives. A single question gets a reply from the phone while they’re standing in a queue.

Once somebody has replied once, they’ll reply again. So you send the second question a few days later and the third after that. You assemble the story across three short exchanges instead of asking for a paragraph nobody has time to write.

Copy-and-paste AI prompt
Here are replies from customers about their experience with my product: [paste everything you have, however rough]. Turn each one into a short story with four parts. Where they were before they bought. What they’d already tried that hadn’t worked. What the first week with the product was like. What’s different now, including any numbers they mention. Keep their own words wherever you can, and tell me which of these stories would land hardest on somebody who hasn’t bought yet.

When the stories come back, read them for the detail you didn’t expect rather than for the result. The unexpected detail is the part a reader will believe, and it’s usually something small that the customer mentioned in passing.

Somebody saying they used to do this at the kitchen table after their kids went to bed tells your reader more than any percentage could. It says the person was ordinary and tired, which is exactly what the reader is.

Part 3 of 7

Choosing Which Number to Use and Which to Leave Out

03

Once you have several results, you have to choose between them, and most people choose badly.

They pick the largest one because the largest one seems most impressive, and impressive isn’t the same thing as persuasive.

The number you want is the one closest to what your reader could plausibly achieve. Not the ceiling, and not the average either, but the result that somebody in their position would reach if things went reasonably well for them.

Reaching for the ceiling costs you more than it gains. A reader who can’t imagine getting your headline result concludes the product isn’t for people like them, and they leave without telling you that’s why.

Averages have a different problem.

An average is a number that nobody in particular ever got, and readers can sense that. It’s a statistic rather than an outcome, and statistics belong in reports, not in writing meant to make somebody feel something.

Ceilings also age badly. A headline result from three years ago starts to look suspicious when nobody has matched it since. Readers can smell a number that has sat on a page too long with no company.

So you want a real person’s real result, sitting at the modest end of what’s possible.

That result does something the ceiling can’t do, which is to make the reader think they could probably manage at least that much.

That thought is quiet and it arrives without fanfare. Nobody announces it to themselves. It just settles, and once it has settled the reader stops evaluating and starts working out whether they can afford it.

Probably managing at least that much is the exact thought that precedes a purchase. Nobody buys because they’re certain of a spectacular outcome. They buy because they believe a modest outcome is within reach and worth the price.

There’s a related error that shows up in bundles and course sales constantly.

Sellers quote the best result their best student achieved, and then they wonder why beginners bounce off the page. The beginner did the arithmetic and concluded they aren’t that person.

You can avoid that with one sentence. Say plainly that the big result belongs to somebody experienced, and then show the modest one and say that’s what a beginner tends to see. Honesty about the range makes both numbers believable at once.

There’s a second question about numbers that people rarely ask themselves, which is whether to use one at all.

Some products are better sold with no figures anywhere, and forcing a number into them makes the whole piece less credible.

If your product improves something hard to measure, like confidence or clarity or how much a person dreads a task, then a number will always feel bolted on. The reader will sense the strain and it will cost you more than the number gains.

The reader will sense the strain and it will cost you more than the number gains.

In those cases the proof is behavioural rather than numerical. Somebody who used to put a task off until Sunday night now does it on Wednesday. That’s evidence, and it’s stronger evidence than a fabricated percentage would ever be.

There’s a version of behavioural proof that works even better, and it involves what a person stopped doing. Somebody who cancelled a subscription they had been paying for since they started has told you something no percentage could carry.

Behavioural proof is available to everybody and almost nobody uses it.

Ask your customers what they do differently now, in terms of their calendar and their habits, and you’ll get evidence that’s impossible to fake and easy to believe.

⚠️
Watch out

Never round a number up to make it sound better. A reader who sees a suspiciously clean figure assumes it was chosen rather than measured. An awkward number reads as true precisely because nobody would have picked it on purpose.

That principle goes further than most people take it. Precise numbers beat round ones every single time. Precision suggests somebody measured, while roundness suggests somebody estimated, and estimation always smells faintly of wishful thinking.

So a customer who went from eleven sales a month to nineteen is more convincing than one who doubled. Doubling sounds like a story somebody told themselves. Eleven to nineteen sounds like a person who counted.

Copy-and-paste AI prompt
Here are all the results I have from customers: [paste them]. My typical buyer is [describe them, including how much experience and time they have]. Sort these results into three groups. First, results a beginner in my audience could plausibly reach. Second, results only an experienced person would reach. Third, results so large they’d make a beginner feel further behind rather than encouraged. Then tell me which single result I should build my main proof around, and why.

Use the first group in everything you write to cold and new buyers, because those are the readers who need to see themselves in the result. Save the second group for people who already trust you and are wondering how far this could go.

The third group is best left out of your marketing entirely, however tempting it’s to include. It only works on people who are already convinced, and people who are already convinced don’t need convincing.

Part 4 of 7

Placing Testimonials Where the Doubt Is

04

The standard way to use testimonials is to gather them into a block, put a heading above it, and place the block somewhere on the page.

That arrangement guarantees they’ll be skipped, and it’s what almost everybody does.

They get skipped because a block of praise has no question attached to it. The reader isn’t asking anything at that moment, so nine answers arrive in response to nothing, and the eye treats them as decoration and passes over.

Testimonials work when they answer a doubt the reader is having at that exact moment. So instead of a block, you place them singly, each one sitting immediately after the sentence that would have raised the doubt it addresses.

You can see the difference in your own behaviour on any sales page you’ve read this year.

The testimonial block gets a glance. A single quote sitting under the exact sentence that worried you gets read word for word.

That requires knowing what the doubts are and in what order they arrive, which is a job you can do in twenty minutes with a piece of paper. Write down every reason somebody wouldn’t buy, in the order those reasons would occur to them.

The list is shorter than you’d think and it’s remarkably consistent across niches.

Will this work for someone like me. Have I already tried this. Is it too much work. Will I be left alone with it. Is it worth the money.

Write those five down and you have the skeleton of the page. Every section you write from now on has a job, and the job is to remove one of them. Nothing goes on the page unless it removes one of the doubts on that list.

Each of those doubts has a testimonial that answers it, and the answers are all different.

The one about whether it works for someone like them needs a customer who started where the reader is standing, with the same disadvantages.

The one about having tried this before needs a customer who says so explicitly. Somebody who mentions the three other things they’d bought and abandoned, and then explains what was different this time, dismantles a doubt no argument can touch.

The one about being left alone afterward needs a testimonial about what happened after the sale.

Nobody collects these, because nobody asks. Ask two customers what happened when they got stuck, and you’ll have the answer to a doubt that costs you sales without ever being spoken.

There’s a doubt nobody puts on their list and it stops more purchases than any of the others. It’s the doubt about whether they’ll follow through. They have bought things before and left them unopened, and they remember every one of them.

They have bought things before and left them unopened, and they remember every one of them.

A testimonial from somebody who admits they usually abandon things, and then says why they didn’t abandon this one, answers a question your reader was too embarrassed to ask. It works precisely because they would never have raised it themselves.

The doubt about price is answered last and answered indirectly. A testimonial that mentions the price is weak, because it sounds coached. A testimonial where somebody says what the problem was costing them settles the price question without mentioning it.

Length works against you here as well.

A testimonial that runs six sentences gets skimmed like everything else, and the sentence that answers the doubt is usually buried in the middle where nobody reaches it.

Placement is what turns each of these from decoration into work. Answer the doubt in the same breath it forms, and the reader’s objection never becomes a decision. They think it, they see it addressed, and they carry on reading.

⚡
Shortcut

Read your page out loud and stop at every point where you’d expect a friend to interrupt with a question. Those interruption points are where your testimonials belong, and there are usually four or five of them on a page.

Use the customer’s full name and something identifying wherever they’ll allow it. An anonymous testimonial is worth almost nothing, because the reader assumes that anything unattributed was written by you at eleven at night when the page was nearly finished.

Trim each testimonial down to the part that answers the doubt, and cut everything else. A long testimonial contains one useful sentence and four sentences of praise, and the praise weakens the useful sentence by surrounding it.

Nobody minds a shortened testimonial and nobody thinks less of you for it.

What they notice is whether the sentence in front of them answers the question in their head, and a trimmed quote does that faster than a full one.

Copy-and-paste AI prompt
Here’s my sales page: [paste it]. Here are my customer testimonials: [paste them all]. First, list every doubt a reader would have while reading this page, in the order those doubts would arise. Second, match each doubt to the testimonial that best answers it, and trim that testimonial down to the one or two sentences that do the answering. Third, tell me exactly where on the page each trimmed testimonial should sit. Flag any doubt I have no testimonial for.

That last flag is the most valuable part of the exercise. The doubt with no testimonial against it’s the one costing you the most sales, and now you know which customer to email and what to ask them for.

Go and get that one specifically. A single testimonial that answers your biggest unaddressed doubt beats twenty general ones. You can usually get it with a two-line email to somebody who already likes you.

✅ Do this
  1. Write down every reason somebody wouldn’t buy, in the order those reasons would occur to them.
  2. Write those five down and you have the skeleton of the page.
  3. Use the customer’s full name and something identifying wherever they’ll allow it.
  4. Go and get that one specifically.
Part 5 of 7

Proof When You Have No Results Yet

05

Everything so far assumes you have customers and results.

Plenty of people don’t, and the usual response is to write around the problem with vague language, which is the worst thing you could do.

The temptation to fudge is strongest in the first month of a launch, when you have nothing to show and a page to fill. That’s exactly the month when a reader is most alert, because a new product with no history is the sort of thing people examine carefully.

Vagueness is louder than an admission.

A reader who meets phrases like proven methods and time-tested principles with no evidence anywhere near them concludes you have nothing. They get there faster than if you had simply told them.

So say so. A new product with no results yet isn’t a shameful thing, and readers are far more forgiving about it than sellers expect. What they won’t forgive is being misled about it.

Readers are used to sellers hiding weaknesses, so a disclosed weakness reads as a signal about your character rather than about your product.

That signal is worth more than the weakness costs you, which is a trade almost nobody is willing to make.

Saying it plainly does something you might not expect. It makes everything else you’ve written more credible, because a person willing to admit the one weak thing is assumed to be telling the truth about the rest.

There’s a version of the admission that goes too far and turns into apology, and you want to avoid that.

Saying you have no results yet is a fact. Saying you hope this works and you’re sorry you can’t promise anything is a confession, and confessions frighten buyers.

State it once, plainly, with no apology attached to it anywhere. Then carry on as though the question has been settled, because for most readers it has been. They wanted to know, you told them, and now they’re listening to what comes next.

Then you replace results with the other three kinds of proof, and all three are available to you on day one.

Your own result, the logic of the method, and the small early signals from the first people who tried it.

Your own result counts, provided you tell it straight and don’t inflate it. If the method worked for you, say what it did, with the starting point included and the timeline stated. One real story beats a hundred manufactured ones.

Your own result counts, provided you tell it straight and don’t inflate it.

The logic of the method is the most underused proof there is. Explain why the thing works, step by step, in enough detail that the reader can check the reasoning themselves. A reader who follows your reasoning and agrees with it has been persuaded.

That kind of persuasion is durable in a way that testimonials aren’t. Somebody who believes your logic stays convinced even when they hit a rough patch. They understood why the thing should work, rather than trusting that it worked for a stranger.

Explaining the logic also does something for you rather than just for the reader.

If you can’t explain why your method works, you may not have a method. Finding that out before you launch is a great deal cheaper than finding it out afterward.

The third kind is early signals, and you can start collecting them within a week of launching. Give the product to five people for nothing, ask them what happened, and report what they tell you without dressing it up.

Give it to people who are close to your typical buyer rather than to your friends.

A friend will be kind and useless. A stranger with the problem will tell you where they got stuck, and where they got stuck is the most valuable thing you can learn before launch.

Even a modest early signal is powerful when it’s honest. Somebody saying they tried it for two days and one small thing worked is more convincing than a polished testimonial, because polished testimonials are what everybody else has.

💡
Tip

Write the sentence admitting you have no results yet, and put it early rather than burying it. Readers who find a weakness you disclosed will trust you. Readers who find one you hid will stop reading and never come back.

There’s an approach that turns the absence into an advantage, and it works better than it has any right to. Invite the reader to be one of the first, and tell them plainly that you want their results because you don’t have enough yet.

That’s an honest request and people respond to it. A buyer who understands they’re helping you build something is a buyer who feels involved, and involved buyers complete the product and send you the story you needed.

Copy-and-paste AI prompt
I have a new product with no customer results yet. It’s about [describe it] and my buyer is [describe them]. Help me build proof without any testimonials. First, walk through the logic of why this method works, step by step, in a way a sceptical reader could check for themselves. Second, tell me how to present my own result without inflating it, including what I should admit about my starting point. Third, write me the sentence that admits I have no customer results, in a way that builds trust rather than costing it.

Read the logic section carefully before you use it, because that’s the one that has to hold up. If your explanation of why the method works has a weak link in it, a sharp reader will find that link, and finding it costs you the sale.

Better to discover the weak link yourself and either strengthen it or leave it out. Nobody expects a method to be flawless. They expect you to know where its edges are, and knowing that makes you sound like somebody who has used it.

Part 6 of 7

Balancing Feeling and Evidence Without Killing Either

06

There’s a worry that stops people doing any of this properly, and it goes like this.

Lean too far into feeling and you sound soft and unserious. Lean too far into evidence and you sound like a spreadsheet, and nobody reads spreadsheets.

The worry assumes the two things compete for the same space, and they don’t. Feeling and evidence do different jobs at different moments, and the trouble only starts when you try to do both at once in the same sentence.

Feeling comes first and it opens the door.

A reader who feels nothing has no reason to evaluate your evidence, because evaluation takes effort and nobody spends effort on something they aren’t already interested in.

That order gets reversed constantly by people who are nervous about seeming unserious. They lead with the credentials and the numbers, and the reader never gets far enough in to care about any of it.

Evidence comes second and it closes the door behind them. A reader who feels something but sees nothing solid will enjoy the writing and then leave, because feeling alone doesn’t justify spending money and they know it.

You can see the two working together in any piece of writing that ever persuaded you of anything.

The feeling made you care enough to look, and then something solid gave you permission to act on the caring.

So the order is fixed, and it repeats. Make them feel it, then give them something to check. Make them feel the next thing, then give them something to check that. The page runs in that rhythm from the top of it to the bottom.

The rhythm is easier to hear than to plan.

Read anything that held your attention all the way to the end, and you’ll find it alternating. Something to feel, then something to check, and then something to feel again.

Getting the rhythm wrong in one direction produces writing that reads as sentimental. All feeling and no substance, which is the trap people fall into when they’ve been told that storytelling is what sells and they’ve taken the advice too literally.

Getting it wrong in the other direction produces writing that reads as a brochure.

All specification and no reason to care, and this is where most people who came from a technical background end up without meaning to.

The proportion that works is roughly two parts feeling to one part evidence, and it will feel light on evidence to you. That’s because you find your evidence interesting, and you’re the only person in the world who does.

That’s because you find your evidence interesting, and you’re the only person in the world who does.

The two-to-one proportion also holds within a single section. Two paragraphs describing what somebody went through, then one paragraph of what it produced. Any tighter than that and the section starts to read like a case study nobody asked for.

Your reader doesn’t find it interesting. They find it necessary, which is a different thing. They need enough to justify a decision they’ve already half made emotionally, and one solid piece will usually do it where five won’t.

That’s the last thing to understand about proof, and it’s the thing that changes how much of it you collect.

You don’t need more proof. You need the right proof, in the right place, attached to a person your reader can see.

🧠
Idea

Go through a finished piece and mark every paragraph as either feeling or evidence. If you see three evidence paragraphs in a row, the reader stopped somewhere in the second one. Break the run with a story and the rest gets read.

Marking the paragraphs takes five minutes and it shows you the shape of the piece in a way that reading it cannot. You’ll see the runs. You’ll see where the evidence piles up, and you’ll see the long stretch with nothing solid in it.

Both problems are easy to correct once you can see them, and neither one is visible while you’re writing. While you’re writing you’re down inside the sentence, not standing above the page and looking at the shape of it.

Copy-and-paste AI prompt
Here’s a piece I’ve written: [paste it]. Go through it paragraph by paragraph and label each one as either feeling or evidence. Show me the sequence as a simple list. Then tell me where the evidence runs too long without a break, and where the writing goes too long with nothing solid in it. For each problem spot, tell me what kind of material I’d need to put there and roughly how long it should be.

Run that on the piece that underperformed, not the one that did well. The piece that failed usually failed for a structural reason rather than a stylistic one, and this will show you the reason in about a minute.

Nine times out of ten you’ll find three evidence paragraphs stacked together somewhere in the middle. That’s where your readers left, and no amount of rewriting the headline was ever going to bring them back.

It helps to remember what a reader is doing when they look at your proof.

They’re frightened, in a small and unglamorous way. They’ve spent money before on something that didn’t work, and they don’t want to feel like that again.

Fear makes people read differently than confidence does. A frightened reader is hunting for a reason to say no, because saying no is safe and costs them nothing at all. Every unsupported claim you make hands them one.

Proof isn’t there to impress that person. It’s there to reassure them, and those two goals pull in opposite directions more often than anybody admits. The impressive number frightens them further. The modest one steadies them.

So the biggest result you have is rarely the most useful thing you own.

The most useful thing is the story of somebody unremarkable who was in the same position, did the ordinary work, and got somewhere they were pleased about.

There’s a habit that makes all of this easier and it costs nothing. Every time a customer says something that makes you feel something, paste it into a file with the date on it. In six months that file will be the most useful thing you own.

That person exists in your customer list right now, and you probably haven’t asked them anything.

They’d tell you if you did. It takes one email and a single question, and the answer is worth more than any headline figure you could put on a page.

Everything else follows from having that story. You place it where the doubt is, you keep the number small and precise, and you let the reader arrive at the belief themselves rather than being handed it.

✅ Do this
  1. Make them feel it, then give them something to check.
  2. Make them feel the next thing, then give them something to check that.
  3. Read anything that held your attention all the way to the end, and you’ll find it alternating.
  4. Run that on the piece that underperformed, not the one that did well.
Part 7 of 7

Put It on Video

07

Here’s the rule that keeps video honest: generated footage can set the scene, but it can never be the proof.

A generated customer, a generated result or a generated screenshot is a fabrication, however lifelike it looks. Use AI scenes for mood and context, then show real evidence on top: a testimonial shared with permission, a genuine screenshot, an actual photo of the finished work.

Time the proof the way you place it on a page. When the voiceover raises the doubt, the evidence should appear on screen within a second or two, held long enough to read on a phone.

Do this on your next video
  1. 1
    List your real results and get permission for each one.
  2. 2
    Note the doubt each result answers.
  3. 3
    Generate only the background scenes.
  4. 4
    Overlay the real proof at the moment of doubt.
Copy-and-paste AI prompt for your video script
Here are my real customer results and testimonials: [paste them]. Plan a 30-second video where generated footage only sets the scene and every claim is backed by one of these real items shown on screen. Tell me which item to show, at which second, and what doubt it answers. Don’t invent any result, number or quote.
Use one idea before the next lesson.

Try it on the next thing you write or film. Then come back for the next lesson.

All video marketing →