John Caples headlines: lessons from tested mail-order ads for your VSL headline and hook
John Caples wrote one of the best-known mail-order headlines of the 1920s and built a career on measuring which ads sold. This guide covers the facts, the headline lessons that transfer to video, original VSL headline and hook examples, and a step-by-step plan for testing headlines on your own traffic.
On this page
- Who was John Caples?
- The story behind "They Laughed When I Sat Down at the Piano"
- What John Caples teaches about headlines
- Applying John Caples headlines to VSL headlines and hooks
- Why Caples tested everything, and how that maps to a VSL
- A headline test plan for your VSL
- What if a headline test is inconclusive?
- Common headline mistakes that testing exposes
- How to apply this to your VSL tomorrow
- Frequently asked questions
John Caples headlines are famous because they were tested. His mail-order ads were judged by the responses readers sent back, not by opinions in a meeting. The lessons for a VSL: put the viewer's own interest in the headline, make the promise specific and believable, let the headline pick out the right viewer, and test variants on real traffic, judging by plays, pitch reach and revenue.
Who was John Caples?
John Caples was born in New York City in 1900 and died in 1990. He attended Columbia University, earned an engineering degree from the U.S. Naval Academy, and began his advertising career at the agency Ruthrauff & Ryan in the 1920s. He joined BBDO in 1927 and rose to vice president (Wikipedia). The Caples Awards site adds that he returned to BBDO as creative director after serving in the Navy during the war, and retired in 1983 (Caples Awards).
His book Tested Advertising Methods first appeared in 1932 and has been revised several times; a revised fifth edition, edited by Fred E. Hahn, was published by Prentice Hall in 1997 (Google Books). He also wrote Making Ads Pay (1957) and How to Make Your Advertising Make Money (1983), and was named to the Copywriters Hall of Fame in 1973. David Ogilvy admired him as a tireless analyzer and teacher of advertising.
The thread through his career was measurement. The Caples Awards site notes that he tested the size, color and position of ads to find what worked best, an approach that not everyone around him liked.
His career, date by date:
John Caples, 1900–1990
His life and books, by date
- 1900Born in New York City
- 1926The piano ad"They Laughed When I Sat Down at the Piano", a mail-order ad for the U.S. School of Music, written in his first year at Ruthrauff & Ryan. Wikipedia dates it to 1926.
- 1927Joins BBDOHe rose to vice president there.
- 1932Tested Advertising Methods first appearsThe book has been revised several times since.
- 1957Publishes Making Ads Pay
- 1973Named to the Copywriters Hall of Fame
- 1983Publishes How to Make Your Advertising Make MoneyHe retired the same year, according to the Caples Awards site.
- 1990Dies
- 1997A revised fifth edition of Tested Advertising MethodsEdited by Fred E. Hahn and published by Prentice Hall.
The story behind "They Laughed When I Sat Down at the Piano"
In his first year at Ruthrauff & Ryan, Caples wrote a mail-order ad for the U.S. School of Music, which Wikipedia dates to 1926. Its headline, "They Laughed When I Sat Down at the Piano", continues with a turn: what happened when the narrator actually started to play.
Why it works, in terms you can reuse:
- It's a story in one line. There's a scene, a person and a moment of tension before you've read a word of body copy.
- The stakes are social. Being doubted in front of a room, then proving everyone wrong, is a feeling most readers recognize.
- It promises a reversal. The second half of the headline tells you the laughter is about to turn. You read on to see the turn.
- The reader is the hero. The ad sold a hope: that you, too, could surprise the room. The course was simply the way to get there.
Notice what it doesn't do. It doesn't lead with the product, the price or the teaching method. It leads with the outcome the buyer wanted, told as a story.
What John Caples teaches about headlines
The publisher's description of the fifth edition of Tested Advertising Methods lists what it covers: writing headlines, making offers, choosing appeals, writing copy, designing layouts and running tests you can afford, all built on what testing showed rather than on guesswork. You'll also find plenty of lists online labeled "Caples headline formulas". Treat them as summaries, and check the wording against the book before you quote anything.
Five lessons from his work that transfer directly to a VSL:
1. Lead with the viewer's interest
The piano ad sold a hope, not a course. Another Caples headline, for Phoenix Mutual Life Insurance, spoke directly to men who wanted to quit work someday: a retirement plan sold as a future the reader already wanted. The ad was later included in the collection The 100 Greatest Advertisements. Above a VSL, the headline should name what the viewer gets, not what the product is.
2. Pick out the right viewer
Both headlines select a reader: someone who wishes they could play, someone who wants to stop working one day. A headline that tries to speak to everyone selects no one. How directly you call out the viewer depends on what they already know, which is the subject of the levels of awareness.
3. Be specific and believable
A promise that sounds too big gets discounted. Specific details, such as a time frame, a number or a recognizable situation, make a claim easier to believe, as long as they're true and documented.
4. Make curiosity serve the promise
Curiosity alone earns a glance. Curiosity attached to a benefit earns the play. The piano headline makes you wonder what happened, but the reason you care is that you'd like it to happen to you.
5. Let tests decide
Caples' career rested on measuring which ads pulled. A headline you love is a hypothesis until buyers vote. The same measure-everything idea runs through Claude Hopkins' 1923 book; see our Scientific Advertising summary.
Applying John Caples headlines to VSL headlines and hooks
A VSL has two headlines. The page headline above the player earns the play or the tap for sound. The spoken hook, roughly the first five to fifteen seconds, earns the next minute. They should make the same promise in different words, and both should pick up the promise of the ad that sent the viewer.
Original examples for four fictional products:
Four approaches, two headlines each
Original examples for fictional products: the page headline above the player, then the first spoken line
Story with a reversal (Fretwise, a guitar app for adults)
Page headline: "At 52 I couldn't play a chord. Last month I played at my daughter's wedding."
First spoken line: "Everyone told me I'd started too late. Here's what happened when I picked up a guitar anyway."
Call out the reader (DryDock, a boot dryer)
Page headline: "For anyone who pulls on cold, damp boots every winter morning"
First spoken line: "If your boots are still wet when you leave the house, this takes two minutes to fix."
Specific how-to (Ledgerly, bookkeeping for freelancers)
Page headline: "How freelancers close their books in 20 minutes a month"
First spoken line: "If your books eat a whole weekend every quarter, you're about to see why they don't have to."
News (PitProbe, a wireless meat thermometer)
Page headline: "New: a thermometer that warns you before the meat overcooks"
First spoken line: "This thermometer tells you when to pull the meat before it's too late, not after."
Two rules keep these honest. First, story headlines must be true: if your presenter didn't live it, don't write it in the first person. Second, every specific (20 minutes, two minutes) must be something you can prove. Testimonials must come from real customers with real experience; the FTC's 2024 final rule bans fake reviews and testimonials, including ones attributed to people who don't exist or who never used the product (FTC). This isn't legal advice.
Why Caples tested everything, and how that maps to a VSL
Mail-order advertising gave Caples something most advertisers of his era lacked: a direct count of responses from each ad. Run two versions of an ad in comparable placements and the replies tell you which one pulled more. He tested the size, color and position of ads as well as their words.
A VSL funnel gives you the same kind of feedback, faster:
| Mail-order testing | VSL equivalent |
|---|---|
| Replies or coupons per ad | CTA clicks and orders per VSL version, traced by click ID |
| Two versions of an ad in comparable placements | A sticky traffic split between two versions of the page |
| Cost per inquiry | Cost per viewer who reaches the pitch |
| Cost per sale | Revenue per visitor and cost per acquisition |
| Size, color and position of the ad | Poster, headline length, player placement on mobile |
The difference is speed. Caples waited for the mail. You can see plays within minutes and pitch reach within days, which makes it tempting to call a test too early. Decide how much traffic you need before you start.
A headline test plan for your VSL
Eight steps, from the first drafts to the next challenger:
A headline test plan for your VSL
Eight steps, in order
1Write 15 to 20 headlines
What happens: Mix approaches: story, call-out, specific how-to, news, question.
Why it matters: Volume matters because your first idea is rarely your best one.
2Screen them with four questions
What happens: Does it promise something the viewer wants? Does it pick out the right viewer? Is it specific and believable? Does it match the ad?
3Pick three or four that are truly different
What happens: Test ideas first, then polish wording later.
Why it matters: Testing two near-identical wordings wastes traffic.
4Run them on the same traffic
What happens: Split new visitors evenly and keep each visitor on one version.
5Choose the primary metric before you start
What happens: Above a muted autoplay player, that's unmute rate. On a click-to-play page, it's play rate.
6Size the test
What happens: As a worked calculation: telling a 55% unmute rate from 60% with 95% confidence and 80% power takes roughly 1,500 visitors per version in a standard two-proportion test.
Why it matters: Revenue differences need far more traffic, so treat a headline win as provisional until revenue per visitor agrees.
7Check the guardrails
Why it matters: A headline that wins plays but loses pitch reach is promising something the video doesn't deliver.
8Keep the winner and test a new idea against it
What happens: Record every result, including the losers.
Why it matters: Losers tell you what your market doesn't care about.
A worked example: three headlines on 1,600 visitors each. A, a call-out headline, gets a 54% unmute rate; B, a story headline, gets 61%; C, a news headline, gets 63%. A is clearly behind. C edges B, but a two-point gap at 1,600 visitors each is within normal noise. Pitch reach settles it: 24% of B's visitors reached the pitch against 19% of C's, a gap that is clear at this sample size, most likely because C's promise doesn't match what the video talks about first. B becomes the control, and the next round tests a new story angle against it.
On the Growth plan and up, RevvyScales can run headline and poster variants inside one VSL: each variant is shown by the weight you set, each viewer keeps seeing the same one, and analytics break the results down per variant (split tests inside a VSL). On any plan, a folder can split traffic between whole versions of the page (folders and A/B splits). For the full method, see VSL headline testing and VSL A/B testing.
Here is a folder split between two versions of one VSL, with the numbers a headline test is read by:
- Play ratePlay rate of each version: the primary metric of a headline test on a click-to-play page.
- Pitch rateThe guardrail. A headline that wins plays but loses pitch reach is promising something the video doesn't deliver.
- RPVRevenue per visitor of each version. Treat a headline win as provisional until it agrees.
- Traffic %New visitors split evenly between the versions; each visitor stays on one.
What if a headline test is inconclusive?
Many headline tests end without a clear winner. That's a result, not a failure: it tells you the variants weren't different enough to matter to your viewers. Three decision rules:
- Keep the control. If no challenger beat it on the traffic you planned, the current headline stays. Don't extend the test again and again hoping for a winner; that's how false winners get crowned.
- Go bolder next round. Swap wording tweaks for a different promise, a different reader or a different approach, such as a story instead of a how-to.
- Be careful with segments. Slicing an inconclusive test into mobile, desktop, age and source until one slice "wins" usually finds noise. If you suspect a real segment difference, say mobile viewers reading a long headline, run a fresh test for that segment.
Try the rules on three situations:
Headline tests: what would you do?
Show the answer
Keep the control and test a bolder idea next round. If no challenger beat it on the traffic you planned, the current headline stays. Don't extend the test again and again hoping for a winner; that's how false winners get crowned. Next round, swap wording tweaks for a different promise, a different reader or a different approach.
Show the answer
It's promising something the video doesn't deliver. A headline that wins plays but loses pitch reach is promising something the video doesn't deliver. Check pitch reach and revenue per visitor before you crown a winner.
Show the answer
A story headline against a how-to headline. Testing two near-identical wordings wastes traffic, and small wording tweaks rarely move results enough to measure. Test different promises first, then polish wording later.
Common headline mistakes that testing exposes
- Clever over clear. A pun the viewer has to decode costs a second they won't give you.
- Curiosity with no payoff. A mystery headline can win the play and lose the viewer at 0:30, when they realize the video isn't about them.
- A promise the video doesn't keep. It wins the headline test and loses on refunds. Judge by revenue, not plays alone.
- Opinion instead of evidence. "The team likes version B" isn't a result.
- Testing commas first. Small wording tweaks rarely move results enough to measure. Test different promises.
- A mismatch with the ad. If the ad promised a quick fix and the headline promises a life story, viewers feel lost. For openings that pick up the ad's promise, see VSL hook examples.
How to apply this to your VSL tomorrow
- Write down the promise your ad makes, then check that your page headline and first spoken line make the same promise.
- Rewrite the page headline to lead with the viewer's outcome, not the product's name.
- Add one specific, true detail to the headline: a time frame, a number or a situation.
- Draft 15 headlines across at least four approaches and screen them with the four questions above.
- Pick the three most different and set up a sticky split on the same traffic.
- Decide the primary metric (unmute or play rate) and the traffic you need before launching.
- After the test, check pitch reach and revenue per visitor before you crown a winner.
- Log the result and line up the next challenger.
Frequently asked questions
What is John Caples' most famous headline?
“They Laughed When I Sat Down at the Piano,” written for a U.S. School of Music mail-order ad in the 1920s; Wikipedia dates it to 1926. The headline tells a story in one line, a moment of social doubt followed by a promised reversal when the narrator starts to play. The Caples Awards site describes it as one of the most famous direct-response ads ever written.
What is Tested Advertising Methods about?
Tested Advertising Methods is John Caples' book on making ads more effective through testing instead of guesswork. It first appeared in 1932 and has been revised several times, including a revised fifth edition edited by Fred E. Hahn and published by Prentice Hall in 1997. It covers headlines, offers, appeals, copy, layouts and affordable ways to test them.
Did John Caples work at BBDO?
Yes. Caples joined BBDO in 1927 and rose to vice president. Before that he worked at Ruthrauff & Ryan, where he wrote the piano ad in his first year. According to the Caples Awards site, he returned to BBDO as creative director after serving in the Navy during the war and retired in 1983.
Do Caples' print headline lessons work for video sales letters?
The principles transfer well: lead with the viewer's interest, be specific and believable, pick out the right viewer and let tests decide. What changes is the format. A VSL has a page headline above the player and a spoken hook in the first seconds, and both need to carry the same promise. Video also gives faster feedback through plays, retention and pitch reach.
How many headlines should I test on a VSL?
Write many and test few. Drafting 15 to 20 headlines gives you real choices; testing three or four genuinely different ones at a time keeps each variant's traffic large enough to read. As a worked example, telling a 55% unmute rate from 60% takes about 1,500 visitors per version, so more variants mean a longer test.
Where can I find John Caples' headline formulas?
The most reliable source is his own book, Tested Advertising Methods, which covers headlines alongside offers, appeals, copy and testing. Many online lists of Caples headline formulas are paraphrases of varying accuracy, so check any formula against the book before you rely on it, and treat every formula as a starting point for a test, not a guarantee.