Paper · Marketing

Hopkins’s 1923 rules and Dry’s 2024 tests converge on specific, checkable ad claims, but Sutherland disputes that reply counts measure an ad’s whole value

Claude Hopkins’s rules and Harry Dry’s tests both ask for claims a reader can picture and check. Rory Sutherland credits copy with value that reply counts miss. We merge the three into a ten-question checklist and apply it to five lines from our home page.

In Motion Automation research team · 3 November 2026

From above at night, one person stands still reading a newspaper at a lit downtown newsstand while people and cars streak past.

Abstract

Readers pick what to read from a headline, and much copy, the words written to sell, spends that moment on claims no reader can picture or check. Three sources give rules against it: Claude Hopkins’s Scientific Advertising (1923), built on counting the replies each version of an ad drew; Harry Dry’s three tests for a line of copy, from a 2024 podcast (can a reader picture it, can a reader prove it true or false, and could no competitor say it); and Rory Sutherland’s talks and interviews from 2009 to 2024, which apply behavioral science, the study of how people decide in practice, to advertising. We located every quote by page or timestamp and compared the three principle by principle. Dry’s second test restates Hopkins’s chapter on being specific: both hold that a specific claim is believed because it could be proved false. They part on distinctiveness, where Hopkins credits a brewer who won with claims any brewer could have made, by stating them first. Sutherland agrees that testing pays, but holds that copy also earns its value by lowering a buyer’s uncertainty and by visible spending that makes a promise believable: value a reply count misses, and that Hopkins dismissed as advertising “for general effect.” We merge the three into a ten-question checklist and apply it to five lines from the home page of In Motion Automation, a Tampa Bay firm that automates calls and follow-ups for local businesses. Dry’s tests mark the headline as too general, and Hopkins’s chapter on specifics supplies the fix. All five lines fail Dry’s third test, and Hopkins’s account of the brewer shows why two of them can stay; Sutherland’s account changes the verdict on two more.

Introduction

Readers choose what to read by its first words. “We pick out what we wish to read by headlines,” Claude Hopkins wrote in 1923 (p. 24)1. The words of an ad, a web page or an email are called copy; the headline is its first line; and a landing page is the page an ad or a link sends a visitor to, built around one action the visitor is asked to take. Copy that loses the reader often fails the same way: it makes a claim the reader cannot picture or check. Hopkins listed “Best in the world” and “Lowest prices in existence” as claims that “leave no impression whatever” (p. 35)1. In 2024 the copywriter Harry Dry read out an ad he had passed that day, “don’t just get a job, change an entire industry,” and asked whether anyone could see it, prove it, or tell it from a rival’s (1:47 to 2:10)2. A specific claim names a number, a person, an object or an act. A checkable claim is one a reader could prove true or false; Dry’s word for it is falsifiable.

Two accounts explain how copy that avoids this failure works. The first comes from direct response advertising, which asks the reader for a reply, such as a coupon mailed back or an order, so that replies can be counted. A keyed ad carries a code, in Hopkins’s day printed on its coupon, that traces each reply to the ad that drew it, so that two versions of an ad can be compared by what each sold (pp. 2 to 3)1. Its modern form is the A/B test, which shows two versions of a page to visitors split at random and counts the replies to each. The second account comes from behavioral science, the study of how people decide in practice. People choose with incomplete information and imperfect trust in the seller (30:42 to 31:09)3, and on this account what a buyer values is what the buyer perceives: “All value is perceived value” (5:03)4. Copy can then add value by changing how an offer is seen, with nothing in the offer changed.

Three sources state these accounts as rules a writer can apply. Hopkins was president of Lord & Thomas, an advertising agency his book’s introduction calls “considered the largest in the world” (Introduction)1. His Scientific Advertising, 21 short chapters published in Chicago in 1923, builds every rule on keyed replies. Dry teaches copywriting and writes the Marketing Examples newsletter, which his site says is “Read by 130k”56. He set out three tests for a line of copy on How I Write, the podcast of the writer David Perell, in July 2024; Perell introduces him as “the first person that I call whenever I need help with copywriting” (0:50)2. Rory Sutherland is vice chairman of Ogilvy in the United Kingdom, where he cofounded a behavioral science practice, and the author of Alchemy7. He began as a copywriter at Ogilvy & Mather Direct in 1988 (6:15)8.

Read side by side, the three agree on much and differ in two places a writer meets on every page. One is distinctiveness: whether a line any competitor could truthfully sign can still sell. The other is measurement: whether copy should be judged only by the replies it draws. Both matter most to a seller with no track record, whose readers have no reputation to lean on and whose pages draw too few visitors to settle a choice by testing.

This paper asks four questions. What does each source say, in its own words, with a page or timestamp a reader can check? Where do the three agree? Where do they conflict, and can a writer act on both sides? What does a checklist merged from all three change when applied to real lines?

In brief, Dry’s second test and Hopkins’s chapter on being specific give the same rule for the same reason, and Dry’s first test matches Hopkins’s advice on pictures that sell. Hopkins’s account of a brewer who won with claims every brewer could have made qualifies Dry’s third test: a common fact told first and exactly can become one seller’s own. Sutherland agrees that testing pays, but credits copy with value that a reply count misses, through lowered uncertainty and visible spending, which Hopkins dismissed. The contributions are:

  • each source’s rules in its own words, with page, chapter or timestamp (Section 3);
  • a principle-by-principle map of where the three agree and conflict, and how a writer can act on both sides of each conflict (Section 4, Figure 1);
  • a ten-question checklist with the source of each question (Section 5, Table 3);
  • the checklist applied to five lines from the home page of a local automation firm, with proposed rewrites (Section 6, Table 4).

Sources and method

We read each source in full on 31 October 2026. For Hopkins, that was a Library of Congress scan of the 1923 edition, all 21 chapters19; quotes give the page number printed in the book. For Dry, it was the caption track of the 71-minute How I Write episode2, three pages of his site10116, his X profile, an X post of 18 March 2024 with its image512, and his Instagram posts13. For Sutherland, it was the transcripts of three TED talks41415, two interviews published by Edge1617, the caption tracks of two video interviews83, a conversation in The Spectator18 and the author note of an essay7. Quotes from caption tracks drop repeated words and fillers (um, uh, okay) and change nothing else; timestamps are minutes and seconds into the video. No legal full text of Sutherland’s book Alchemy was available, so no claim here rests on it. A principle enters the comparison when at least two sources address the same choice. The applied example uses five lines from the home page of In Motion Automation as it read on 3 November 202619.

What each source says

Dry: three tests for every sentence

Dry states the tests in the episode’s first two minutes, after Perell asks about the three rules he applies “to every sentence” (1:28)2:

“Can I visualize it? Can I falsify it? Can nobody else say this? So, you get three no’s, you’re probably writing a lot of rubbish. You get three yes’s, you’re onto something.” (1:34)2

He scores two ads against them. The recruitment ad fails all three: “you close your eyes, can you see that? No”; it cannot be proved true or false; and “I could start a recruitment company tomorrow and write” it (1:55 to 2:07)2. New Balance’s “Worn by supermodels in London and dads in Ohio” passes all three: a reader can picture “the dad in Ohio with a barbecue”; “supermodels have worn New Balance”; and “a load of dads in Ohio aren’t going to wear a Prada and a load of supermodels aren’t really wearing Reebok” (2:14 to 2:42)2.

Test 1: can a reader picture it? “If you can’t visualize it, you won’t remember it” (2:58)2. The episode opens with Dry reading a list of phrases (seamless transition, charging pitbull, muscular Irishman, better way, leg of lamb); Perell recalls the three he can picture (0:03 to 0:09)2. “What abstract is is it’s intangible. You can’t drop it on your foot” (3:41)2. Dry’s method is to write the abstract word down and keep asking what it means until an object appears: an app for regaining fitness becomes one for getting off the couch and running 5K, “Couch to 5K,” which Dry calls “the most downloaded, the most popular fitness program of all time” (4:54 to 5:34)2.

Test 2: can a reader prove it true or false? “When you write a sentence which is true or false, it’s like as a writer you’re putting your head on the chopping block” (5:58)2. His method, which he credits to a Reddit thread by a since-deleted account, is “you can’t talk, you can only point” (6:38 to 6:45)2. Describing a friend for a blind date, “he’s funny” is talk; “looks like Ryan Gosling, 6’2"” is pointing (7:47)2. Pointing “gets you off the adjective trail” (8:45)2. Later in the episode he puts the same test as a request: “when you say a fact, it guarantees that you say something” (46:10)2, and “If in doubt, give me a fact” (47:03)2.

Test 3: could nobody else say it? “Never write an ad a competitor can sign” (9:31)2, a line Dry credits, unsure of the name, to “Jim Durkey” (9:26)2. The test “forces you to look a little bit deeper at what you’re selling” (9:37)2. His example is a car ad that he introduces as Volkswagen’s and whose text sells Volvo: “Your car has five numbers on the speedometer. Volvo has six. One could get the impression that the people who made your car lack a little confidence.” “No one else can say that” (9:42 to 9:50)2.

Earlier and related statements. The episode, published on 17 July 2024, is the earliest dated statement of all three tests found; Dry announced it on Instagram the same day213. A narrower form appears in an X post of 18 March 2024, “Good copy is falsifiable,” whose image pairs a crossed-out mock ad, “A drink to unlock your brain’s superpowers,” with a ticked one, “One capsule equals six hours of focus”12. An undated page on his site, which refers to “the 2019 human mind,” already says: “Landing page copy is full of unfalsifiable, blanket claims: ‘more, easier, faster …’ If you want to stand out get specific. You can’t bullshit specifics”11. The same page advises “Call out the type of customer you serve.” His list of tips opens with “Write with your eraser,” warns against “‘landing page words’: Unlock, unleash, enhance, exceed, empower, supercharge, etc.,” and ends: “Copywriting is selling. Don’t romanticize it. The goal isn’t to be clever or cute. The goal is to inspire action” (tips 1, 6, 17)10.

Hopkins: rules from counted replies

Hopkins opens with his method: “This is done through keyed advertising, by traced returns, largely by the use of coupons. We compare one way with many others,” and “When one method invariably proves best, that method becomes a fixed principle” (pp. 2 to 3)1. The rest of the book states the principles his agency had fixed that way. He also defines what advertising is not for: “It is not for general effect. It is not to keep your name before the people” (p. 7)1; this paper uses his phrase, general effect, for advertising judged by anything other than the sales it brings. Table 1 gives the principles that bear on writing a line.

Table 1. Hopkins’s principles for a line of copy, in his words, from Scientific Advertising1. Page numbers are those printed in the 1923 edition.
ChapterHopkins’s wordsPageWhat it asks of a line
2. Just salesmanship“Advertising is salesmanship.” “The only purpose of advertising is to make sales.”7Judge a line by whether it sells.
“Ask yourself, ‘Would this help a salesman sell the goods?’ ”9Write what a good salesperson would say.
“Ads are not written to entertain.” “Don’t think of people in the mass.” “Think of a typical individual, man or woman, who is likely to want what you sell.”10Write to one likely buyer.
3. Offer service“The best ads ask no one to buy.” “They offer wanted information. They cite advantages to users.”13Lead with what the reader gets.
“People can be coaxed but not driven.”15
4. Mail order advertisingFor one five-dollar article, replies to one ad cost 85 cents each, to another $14.20, and to a third 41 cents.17 to 18Know what each reply costs.
“In mail order advertising the pictures are always to the point. They are salesmen in themselves.”19Show what sells, not decoration.
“The more you tell the more you sell.”20Give the whole case.
5. Headlines“The purpose of a headline is to pick out people you can interest.”23Make the right reader stop.
“It is not uncommon for a change in headlines to multiply returns from five to ten times over.”25Test headlines first.
“Address the people you seek, and them only.”27
6. PsychologyPuffed Wheat and Puffed Rice “were made successful largely through curiosity”: “Foods shot from guns.” “We learn that people judge largely by price.”29
“An offer limited to a certain class of people is far more effective than a general offer.”32Limit the offer to a named group.
“An identical offer made in a different way may bring multiplied returns.”34Try other framings of one offer.
7. Being specific“Platitudes and generalities roll off the human understanding like water from a duck.” “A man who makes a specific claim is either telling the truth or a lie.”35Make claims a reader could check.
A mail order seller of women’s clothing replaced his slogan, “Lowest prices in America,” once rivals copied it, with “Our net profit is 3 per cent.”36Replace a superlative with a figure.
A brewer described a plate glass cooling room, bottles washed four times, wells sunk 4,000 feet: “All the claims were such as any brewer might have made.” “But he was the first to tell the people about them.”37 to 38A common fact, told first and exactly, becomes yours.
8. Tell your full story“When you once get a person’s attention, then is the time to accomplish all you ever hope with him.”39Put every important claim in every ad.
“Take the opinion of nobody, the verdict of nobody, who knows nothing about his returns.”42Judge by counts, not opinions.
11. Information“Impressive claims are made far more impressive by making them exact.” A drink’s food value became “425 calories per pint. One pint is equal to six eggs in calories of nutriment.”58Research until one exact fact appears.
12. Strategy“There is nearly always something impressive which others have not told.”61
13. Use of samples“The product itself should be its own best salesman.”65Let interested readers try it.
“Give samples to interested people only.”70
15. Test campaigns“Almost any question can be answered, cheaply, quickly and finally, by a test campaign.” “Now we let the thousands decide what the millions will do.”77 to 78Try small and count before spending big.
17. Individuality“When we claim an improvement, naming the man who made it adds effect.”88Name a person behind a claim.
18. Negative advertising“Tell people what to do, not what to avoid.” “Don’t show the wrinkles you propose to remove, but the face as it will appear.”89Show the result the buyer wants.
In his tests “the positive ad outpulls the other four to one.”90
19. Letter writingSuccessful letters “place a limit on an offer. It expires on a certain date.” “Strike while the iron is hot.”93 to 94Give a dated reason to act now.

Sutherland: value as the buyer perceives it

Sutherland’s sources are talks and interviews, so his rules come as claims with cases rather than as a system. Three terms carry them. Reframing is changing how people see a thing without changing the thing: “the power of reframing things cannot be overstated” (1:18)15. Satisficing is choosing something good enough and unlikely to be bad instead of searching for the best: “we describe our behavior as if we’re maximizing but most of the time we go ‘I want something that’s pretty good and definitely isn’t awful’”16. A costly signal is a message believed because it was costly to send, as a peacock’s tail is costly to grow: “humans, like peahens, attach significance to a piece of communication in some way proportionally to the cost of generating or transmitting it”17. Table 2 gives his principles that bear on copy.

Table 2. Sutherland’s principles that bear on copy, in his words. Timestamps are minutes and seconds into the video; the Edge interviews and the Spectator conversation are text.
PrincipleSutherland’s wordsSource
Perceived value“All value is perceived value.” Relaunching a “square, whole-grain cereal” as “Diamond Shreddies,” with no change to the cereal, is “the most perfect case of creating intangible, added value, without changing the product in the slightest.”Sutherland (2009), 5:03, 11:43 to 12:454
ReframingA pilot, unable to reach a gate: “The good news is that the bus will take you right next to passport control, so you won’t have far to walk with your bags.”Sutherland (2017)17
Lowering uncertainty“The single best improvement in passenger satisfaction on the London Underground, per pound spent” came from dot matrix boards on the platforms, “because the nature of a wait is not just dependent on its numerical quality, its duration, but on the level of uncertainty you experience during that wait.”Sutherland (2011), 7:5115
Uber’s map: “humans have a great need for certainty,” and the map supplied it.NEXT with John Koetsier (2020), 17:08 to 17:198
An American Express application form made its shortness a reason to apply: “the reason this form is so short is because we already know we want you.”NEXT with John Koetsier (2020), 8:58 to 9:27, 11:348
Satisficing and trust“One of the reasons people pay a premium for brands isn’t because they’re objectively better, it’s because they’re less likely to be terrible.”Sutherland (2014)16
Recounting David Ogilvy and Joel Raphaelson, people “bought brand A over brand B not because they thought it was better, but because they were more certain it was good.”Sutherland (2017)17
Costly signallingAdvertising “is a reliable ‘costly signal’ because you wouldn’t do it if you had a bad product.” “When you make it more efficient and less costly, you make it less convincing.” Cheaper, targeted media “may appear to be efficient, but it’s less effective.”Sutherland (2017)17
Small changes“If you actually want people to remember you and to appreciate what you do, the most potent things are actually very, very small.” A government program can gain uptake “simply by redesigning the form and writing it in comprehensible English.”Sutherland (2010)14
Missing factsIn direct marketing, an ad could explain price and benefits, but if it left out “allow 28 days for delivery” or when a product would arrive, sales fell.NEXT with John Koetsier (2020), 7:50 to 8:108
A weakness reframed“Avis is number two in rental cars” is “an ad for Hertz standing on its own”; add “So we try harder” and “it’s an ad for Avis.”NEXT with John Koetsier (2020), 14:15 to 14:248
Testing“I’m a huge believer in testing.” In direct marketing tests “ostensibly trivial changes in phraseology or presentation would have an enormous effect.”NEXT with John Koetsier (2020), 5:59, 12:078
“If you confine yourself only to testing the things that you think should be important, you’re narrowing your solution space dangerously.”NEXT with John Koetsier (2020), 12:56 to 13:058
A finding from advertising: “the more you test, the more successful you are. There are no diminishing returns.”Pinker and Sutherland (2021)18
“The opposite of a good idea can be another good idea in marketing; that’s not true in physics.”NEXT with John Koetsier (2020), 30:268
What a count misses“What people who are trying to be scientific do is they measure the objective reality, whereas it’s the subjective experience which is really decisive.”NEXT with John Koetsier (2020), 17:538
A limit“You can never get away with being rubbish simply by being [a] good marketer.”NEXT with John Koetsier (2020), 18:128

Where the three agree and where they conflict

Figure 1 sets the principles side by side. On a specific claim, Hopkins’s seventh chapter and Dry’s second test say the same thing; on distinctiveness and on spending to be believed, the sources are in open conflict.

Figure 1. Each principle compared in this paper, and what each source says about it. A dark cell marks a source that states the principle; a light cell, one that states a related or partial form; a red cell, one that limits it or argues against it; a white cell, a principle not found in the sources read for that author.

A specific claim is believed because it could be false

Dry’s second test restates Hopkins’s seventh chapter in nearly the same terms. Dry: a checkable sentence puts “your head on the chopping block” (5:58)2. Hopkins: “A man who makes a specific claim is either telling the truth or a lie,” and “People do not expect an advertiser to lie” (p. 35)1. Both locate the force of a specific claim in the risk of making it. Hopkins adds a method Dry’s “If in doubt, give me a fact” implies: research until one exact figure appears, as when a drink’s food value became six eggs a pint (p. 58)1. The sources read for Sutherland state no rule on specificity, though his cases turn on specific objects: a countdown board, a map, a form that says why it is short.

A picture a reader can see

Dry’s first test asks for words a reader can picture, because “If you can’t visualize it, you won’t remember it” (2:58)2. Hopkins writes about pictures printed in the ad, and makes the same demand of them: in mail order “the pictures are always to the point. They are salesmen in themselves” (p. 19)1, and an ad should show “the face as it will appear” (p. 89)1. Both ask that what the reader sees be the thing sold or its result.

Distinctiveness: a line no rival could sign, or a fact told first

Dry’s third test rejects any line a competitor could sign (9:31)2. Hopkins tells of a brewer who pictured his plate glass cooling room and his 4,000-foot wells when every brewer advertised “Pure.” “All the claims were such as any brewer might have made,” he writes. “But he was the first to tell the people about them” (pp. 37 to 38)1. Hopkins generalizes: “there is nearly always something impressive which others have not told” (p. 61)1. Sutherland’s Avis case gives a third route, in which a fact a rival would never state, being second, becomes a reason to choose the seller once reframed (14:15 to 14:24)8.

The conflict narrows when the test is asked of what rivals in fact say rather than of what they could say. A writer can run Dry’s test against the competitor pages a buyer will see; a line that passes there is distinct in Hopkins’s sense, even if a rival could copy it later. Hopkins’s case suggests the copy holds its value after rivals follow, because the first to say it exactly is the one the reader remembers saying it, though the book reports no count for that case.

How much to say

Dry’s first tip is “Write with your eraser” (tip 1)10; Hopkins writes “The more you tell the more you sell” (p. 20)1. The two work at different scales. Dry trims words inside a sentence; Hopkins asks that the ad carry every fact a buyer needs to act, because the reader may see only this one. Sutherland sides with Hopkins on missing facts: an ad that left out when the product would arrive lost sales (7:50 to 8:10)8. A writer can follow both, with short sentences and a complete case.

Measurement: what a reply count can trace

Measurement is where the sources conflict most. Hopkins built his book on counting replies, told readers to “Take the opinion of nobody, the verdict of nobody, who knows nothing about his returns” (p. 42)1, and ruled out advertising “for general effect” (p. 7)1. Sutherland agrees that testing pays. He learned in direct marketing that small changes in wording had large effects, and he holds that more testing brings more success (5:59, 12:07)818. He parts from Hopkins on what to leave out of the count. Costly signalling holds that money spent where everyone can see it makes a promise believable, so that cheaper, targeted media “may appear to be efficient, but it’s less effective”17. Satisficing holds that buyers pay a premium for a known name because it is “less likely to be terrible”16. Both are value a keyed reply cannot trace to one ad, and both are what Hopkins called general effect. Sutherland’s summary is that people trying to be scientific “measure the objective reality, whereas it’s the subjective experience which is really decisive” (17:53)8.

A writer can act on both. Count every reply that can be keyed, as Hopkins says. Then keep copy whose work is to make a reader trust the seller, such as a named maker, a guarantee or a visible commitment, even where no key can reach it, because Sutherland’s account predicts that cutting it lowers replies the count will attribute elsewhere.

What to test

Hopkins tested to find the one best method and fix it as a law (p. 3)1. Sutherland holds that “the opposite of a good idea can be another good idea in marketing” (30:26)8, so a test that tries only what seems logical narrows the search (12:56 to 13:05)8. The two agree that ideas should be settled by test rather than argument; they differ on which ideas deserve a test.

A checklist for an ad, a headline or a landing page line

Table 3 merges the three sources into ten questions to ask of a draft line. A line need not pass all ten; each failed question names a rewrite to try.

Table 3. Ten questions for a line of copy, each with its source.
#QuestionA yes looks likeSource
1Who is it for?The line, or the line above it, names the reader it wants.Hopkins, ch. 5, p. 23; ch. 6, p. 32; Dry’s site
2Can a reader picture it?It names a person, object, number or scene, not an abstract noun such as “a better way.”Dry, test 1
3Can a reader prove it true or false?A number, a name or an act; it points instead of describing.Dry, test 2; Hopkins, ch. 7, p. 35
4Could a competitor sign it?Put a rival’s name on it. If it still reads true, find the detail only you can state, or state a common true detail first and exactly.Dry, test 3; Hopkins, ch. 7, pp. 37 to 38
5Does it offer something instead of asking for the sale?Information, a result or a sample.Hopkins, ch. 3, p. 13; ch. 13, p. 65
6Does it lower a worry?It says what happens next and when: the date, the wait, the refund, why the form is short.Sutherland, Uber map and short form; Hopkins, ch. 19, p. 93
7Does the page tell the whole story the buyer needs to act?Price, terms, delivery date, who built it.Hopkins, ch. 4, p. 20; ch. 8, p. 39; Sutherland, missing facts
8Is it said in plain, positive words a salesperson would use?No slogans, no “unlock” or “empower,” no warning where a result would do.Hopkins, ch. 2, p. 9; ch. 18, p. 89; Dry, tip 6
9Would another frame change its value with the product unchanged?A weakness stated as a reason, or the same fact against a different comparison.Sutherland, reframing; Hopkins, ch. 6, p. 34
10How will you count it, and what will the count miss?Each version carries its own key (a source tag, a page, a code); the odd version is tested beside the logical one; copy that builds trust stays where no key reaches it.Hopkins, ch. 1, pp. 2 to 3; ch. 15, p. 77; Sutherland, testing and costly signalling

Applied example: five lines from a local automation firm’s home page

In Motion Automation, which publishes this paper, builds automation for small local businesses in Tampa Bay, Florida: software that does a step of the work without a person, such as answering a call, replying to a new lead or following up on a quote. A lead is a person who has asked a business about buying, by phone, web form or email; a quote is the price a business sends for a job before the customer agrees to it. The firm’s home page sells to owners of law firms, insurance agencies, medical and dental practices, and heating, cooling, plumbing and roofing businesses19. Two terms appear in the lines below. An AI receptionist is software built on artificial intelligence (AI) that answers a business’s phone in a natural voice, takes the caller’s details and books the job. The audit is a paid first step, described on the page as “We go through your call logs, inbox and quotes and show you, in dollars, what slips through each month”19.

Table 4 applies the checklist to five lines as the page read on 3 November 2026. Dry’s third test is scored as question 4 of the checklist puts it: with a rival’s name on the line, does it still read true? Every rewrite is proposed; none has been tested with readers.

Table 4. Five home page lines against Dry’s tests (picture, check, only us), Hopkins and Sutherland. Lines as worded on 3 November 202619. Rewrites are proposed.
LineWording reviewedDry’s testsHopkins; SutherlandVerdict and proposed rewrite
Home page headline“Handle more clients with automation.”Picture: no (no person, object or number). Check: no (“more” gives no count). Only us: no; any automation firm could sign it.The line above it, “Automation leaders for Tampa businesses,” names a place but no kind of reader (ch. 5); “more clients” is a generality (ch. 7). Sutherland: it names a gain and leaves out the worry an owner brings, work lost to calls nobody answers.Rewrite. Proposed: “Answer every call, even the one at 7:40 in the evening.” The hour comes from a scene lower on the same page: “Imagine answering every call, even the one at 7:40 in the evening while you’re having dinner with friends.”
Subline“We find bottlenecks that cost you money, like missed calls, unchased leads, forgotten follow-ups and orders typed in by hand. Then we automate them.”Picture: yes (a missed call, an order typed in by hand). Check: partly (no amount for “cost you money”). Only us: no.Names what the reader gets (ch. 3) and each loss by name (ch. 7). Sutherland: “Then we automate them” leaves the reader asking what the fix will cost and when the price is known, a missing fact of the kind that lost sales in direct marketing.Rewrite; Sutherland’s account changes the verdict. Proposed second sentence, in the page’s own terms: “Then we automate them, for a fixed price agreed after the audit and before we start.”
Promise under the main button“If automation won’t pay for itself, we’ll tell you.”Picture: partly (an owner told no; no object). Check: no; a reader learns whether it holds only by asking. Only us: no.“The best ads ask no one to buy” (ch. 3, p. 13). Sutherland: buyers satisfice toward what is less likely to be terrible, and a seller who says it will turn work away lowers that risk even unchecked.Keep the claim; Sutherland’s account changes the verdict. Proposed: “If automation won’t pay for itself, we’ll tell you on the free call,” which names when the reader will hear it.
Service line“Every call answered, even after hours”; under it, “An AI receptionist picks up in your company’s name when your team can’t, takes the caller’s details and books the job on your calendar.”Picture: yes (a call picked up, a job on a calendar). Check: yes; a reader can call after hours and see. Only us: no; any AI receptionist service could sign it.Shows the result the buyer wants (ch. 18); a common service stated first and exactly can become one seller’s own (ch. 7, pp. 37 to 38). Sutherland: “in your company’s name” answers an owner’s worry about how callers are treated.Keep. Proposed addition: a number a reader can call after hours to hear the receptionist, a sample in Hopkins’s sense (ch. 13).
Figure on quotes“Forgotten follow-ups”: “Across 794 contractors, the typical one wins 38% of the quotes they send, but 74% of the ones where the customer gives an answer.”Picture: partly (a contractor’s quotes, no scene). Check: yes (794, 38%, 74%, a named source). Only us: no; any provider could cite the same survey.Exact figures make a claim more impressive (ch. 11); a common fact told first and exactly becomes yours (ch. 7, pp. 37 to 38). Sutherland: it reframes an unanswered quote as work still to be won.Keep; Hopkins’s brewer case outweighs Dry’s third test here. Proposed: follow it with the act it implies, in the page’s own words, “Open quotes get chased until you hear yes or no.”

Dry’s tests mark the headline as too general, and Hopkins’s seventh chapter supplies the fix: name the call and the hour, as the scene lower on the page already does. All five lines fail Dry’s third test, because another automation firm could sign each of them. For the service line and the figure on quotes, Hopkins’s account of the brewer gives the reason to keep them: a common fact told first and exactly can become one seller’s own. Sutherland’s account changes two verdicts that the other tests reach on narrower grounds. The subline names losses a reader can picture but leaves a worry open, what the fix will cost and when the price is set, so the rewrite adds the next step. The promise under the main button fails Dry’s second test but does a different job, lowering the risk an owner feels in paying for automation that may not pay, so it stays.

Discussion

The agreement between Dry and Hopkins has a plain mechanism. A claim a reader could prove false is risky to make, and a reader who sees the risk taken is more inclined to believe the claim. Hopkins says so directly: people “know that he can’t lie in the best mediums” (p. 35)1. Read this way, a checkable claim is a costly signal in Sutherland’s sense, paid for in exposure rather than money, and the sources’ disagreement shrinks to a question of which costs a reader can see. A precise figure shows its cost on the page; a national campaign or a known name shows it only to a reader who already knows what such things cost.

The comparison has limits. Hopkins’s rules rest on his agency’s tests, which the book reports as conclusions without the counts behind most of them. Sutherland’s claims come from talks and interviews, where cases are told from memory, and his book Alchemy, which states them at length, was not read. Dry’s three tests are stated in full in a conversation, and the earliest dated statement of all three is the episode of 17 July 2024; the narrower “Good copy is falsifiable” is dated 18 March 2024, and an undated page of his site makes the same point. Dry credits the line “Never write an ad a competitor can sign” to a “Jim Durkey” whose name he was unsure of, and Sutherland gives no origin for “the opposite of a good idea can be another good idea”; neither was traced. Quotes from video come from automatic captions, which can mishear words; the speakers’ meaning was checked against the surrounding passage. The applied example covers five lines from one firm’s home page, and Dry’s third test was judged by putting a rival’s name on each line, not against captured rival pages.

The checklist suggests tests that could be run. Whether a headline that names a call and an hour draws more booked calls from owners than “Handle more clients with automation” could be tested with two keyed versions of the home page once it draws enough visitors. Whether the promise to say when automation will not pay raises trust beyond a plain offer of a free call could be tested with message tests on owners before any traffic split. And whether a common fact told first holds its value after rivals copy it, as Hopkins’s brewer suggests, would need a market where the same claim can be followed over time.

Sources

  1. Claude C. Hopkins, "Scientific advertising," Lord & Thomas, 1923. loc.gov
  2. David Perell, "Learn copywriting in 76 minutes - Harry Dry," YouTube, 2024. youtube.com
  3. Lenny's Podcast, "What most people miss about marketing | Rory Sutherland (Vice Chairman of Ogilvy UK, author)," YouTube, 2024. youtube.com
  4. Rory Sutherland, "Life lessons from an ad man," TED, 2009. ted.com
  5. Harry Dry, "Harry Dry (@harrydry)," X, n.d.. x.com
  6. Harry Dry, "Copywriting," Marketing Examples, n.d.. marketingexamples.com
  7. Rory Sutherland, "Is everything BS?," Behavioral Scientist, 2024. behavioralscientist.org
  8. NEXT with John Koetsier, "Dark arts & magic in marketing: Rory Sutherland on data and when NOT to use it," YouTube, 2020. youtube.com
  9. Library of Congress, "Scientific advertising, by Claude C. Hopkins, Chicago 1923" (digital scan), McGrinsey, n.d.. mcgrinsey.com
  10. Harry Dry, "17 tips for great copywriting," Marketing Examples, n.d.. marketingexamples.com
  11. Harry Dry, "7 practical ways to write copy that converts," Marketing Examples, n.d.. marketingexamples.com
  12. Harry Dry, "Good copy is falsifiable," X, 2024. x.com
  13. Harry Dry, "Had a buzz cut. Went on my favourite podcast," Instagram, 2024. instagram.com
  14. Rory Sutherland, "Sweat the small stuff," TED, 2010. ted.com
  15. Rory Sutherland, "Perspective is everything," TED, 2011. ted.com
  16. Rory Sutherland, "This thing for which we have no name," Edge, 2014. edge.org
  17. Rory Sutherland, "Things to hang on your mental mug tree," Edge, 2017. edge.org
  18. Steven Pinker and Rory Sutherland, "Sense and sensibility: Steven Pinker and Rory Sutherland on reason vs instinct," The Spectator, 2021. spectator.co.uk
  19. In Motion Automation, "Handle more clients with automation," In Motion Automation, 2026. agentsinmotion.ai

Want us to look at your business this way?

We measure where calls, leads and paperwork slip through, in your own records, before we build anything.

Book a free call
ResearchDownload the PDF14 pages