John Caples Proved This in 1926. Your Subject Line Still Gets It Wrong.
In 1926, John Caples wrote one of the most famous headlines in the history of advertising:
“They laughed when I sat down at the piano. But when I started to play…”
That headline, for a mail-order piano course, ran for years. It generated enormous response. And Caples spent the next five decades figuring out exactly why it worked, testing headlines obsessively at BBDO, the advertising agency where he worked for most of his career.
His book, Tested Advertising Methods, is the most rigorous study of what makes people read advertising that has ever been published. Not the most famous. Not the most widely assigned. The most rigorous. Every claim in it is backed by actual split-test results, run at scale, over decades.
Caples was not a theorist. He was an experimenter. And what he found about headlines applies word-for-word to cold email subject lines.
The headline is the whole game
Caples’s core finding: the headline is responsible for the majority of an ad’s performance. Change the headline on the same body copy and you can change response rates by 500% or more.
He documented cases where two ads with identical body copy, different only in the headline, produced wildly different results. The headline was the variable that mattered most. Everything else was secondary.
“The most important element in an advertisement is the headline. It is the telegram which decides whether the reader will read the copy.”
In cold email, the subject line is the headline. Not in a metaphorical sense. Functionally. It determines whether the email gets opened. An email that isn’t opened doesn’t exist. The most brilliant copy in the world, the most specific proof point, the most perfectly timed ask, all of it is invisible if the subject line failed.
Yet subject lines are consistently the least tested part of most cold email programs. Teams spend hours rewriting body copy that 12% of recipients are reading and ignore the subject line that determines whether 12% ever becomes 35%.
Caples would call this working on the wrong variable.
The three types of headlines that work
Through decades of testing, Caples identified three headline types that consistently outperformed all others.
Type 1: Self-interest
The headline directly promises something the reader wants.
Caples considered this the most powerful headline type. Not clever. Not indirect. Directly addressing what the reader gets.
“How to win friends and influence people” is self-interest. “How I improved my memory in one evening” is self-interest. “Retire at 55 with more money than you ever dreamed possible” is self-interest.
Applied to cold email subject lines:
how to book more meetings from cold emailget your emails out of spamscale outbound without hiring more SDRs
These work because they answer the question the reader is constantly, implicitly asking: “What’s in this for me?” They don’t require the prospect to figure out why this is relevant. The relevance is stated directly.
The objection to self-interest subject lines in cold email is that they can sound like advertising. That’s a fair concern. The solution is specificity: “how to book more meetings” sounds generic. “How we booked 41 meetings in 90 days without buying more licenses” sounds like something worth opening.
Type 2: News
The headline announces something new.
Caples found that news headlines performed very well, particularly when the “news” was genuinely relevant to the reader. The word “new” itself had above-average performance in his tests, as did words like “announcing,” “introducing,” and “now.”
Applied to cold email subject lines, news can mean:
- Something new about your product or approach:
new: isolated sending infrastructure for cold email - Something new in the world that’s relevant to them:
new research on cold email reply rates in 2026 - Something new happening at their company that you’re referencing: the subject line itself isn’t “news” but the email references a news event about them
The news type requires something genuinely new to report. If the “news” is recycled or fabricated, it backfires. Prospects are good at detecting fake news, particularly in cold email where they’re already skeptical.
Type 3: Curiosity
The headline creates a gap between what the reader knows and what they feel they need to know.
“They laughed when I sat down at the piano” works because it creates an instant question: why were they laughing, and what happened next? The reader can’t not find out.
Caples was careful to distinguish between curiosity that works and curiosity that doesn’t. Good curiosity is connected to self-interest. The reader wants to know the answer because the answer will benefit them. Bad curiosity is disconnected from the reader’s real concerns. The answer is interesting but not relevant.
In cold email, curiosity without self-interest is the pattern behind clickbait subject lines that get opens and zero replies. “You won’t believe what we found…” gets opened. Nothing in the email justifies the open, and the reader feels manipulated.
Applied well, curiosity in cold email subject lines looks like:
why your cold emails stopped working(curiosity plus pain recognition)the deliverability issue most SaaS teams don't notice until it's too late(curiosity plus stakes)something we spotted on [Company]'s outbound setup(curiosity plus personalization signal)
The curiosity has to pay off. If the email content doesn’t deliver something genuinely worth knowing, the curiosity open just creates a negative brand impression.
The subject lines Caples would never write
Caples was equally rigorous about what didn’t work, and his findings here map almost perfectly onto the worst cold email subject lines in circulation.
Cleverness at the expense of clarity. Caples tested clever headlines against clear headlines repeatedly. Clear won. Every time. “A mysterious way to get more meetings from cold email” is weaker than “get more meetings from cold email.” The mystery adds nothing. The clarity removes friction.
Negative framings. “Don’t make these cold email mistakes” underperforms “how to avoid the cold email mistakes that kill reply rates.” The construction is similar but the second version leads with what the reader gets rather than what they might be doing wrong. Caples found that positive framings consistently outperformed negative ones.
Generic curiosity gaps. “This might surprise you” performs poorly because the promise of surprise is not connected to anything the reader values. “The one thing killing your cold email reply rates might surprise you” is better because the surprise is specific and the stakes are clear.
Subject lines that reference the sender, not the reader. “Introducing BuyerBrains” is a subject line about you. “Scale outbound without more headcount” is a subject line about them. Caples was unambiguous: the reader cares about themselves, not about you. Write subject lines that are entirely about them.
Fake urgency without substance. “Last chance” and “limited time offer” perform well in certain contexts, specifically when there is a real deadline and a real offer. In cold email, where the urgency is manufactured and the prospect knows it, these phrases signal desperation or deception. Skip them.
The “how I” and “how to” construction
One of Caples’s most tested findings: the “how I” and “how to” constructions consistently outperformed other headline structures.
“How I improved my memory in one evening” was tested against dozens of alternatives and won. “How to” subject lines work because they promise a specific, actionable benefit without requiring the reader to already believe a claim.
Compare:
- “Our cold email program gets exceptional results” (a claim requiring belief)
- “How to get consistent results from cold email” (a promise of useful information)
The second version doesn’t require the prospect to believe anything. It just promises to show them something. The barrier to opening is lower.
Applied to cold email subject lines:
how to set up a cold email program that doesn't die after 30 dayshow we fixed a client's deliverability in two weekshow to tell if your cold emails are landing in spamhow I booked 41 meetings from one 200-person list
The “how I” version has an added advantage: it implies a specific, real story. Not a theoretical framework. An actual thing that happened to an actual person. This signals authenticity, which cold email desperately needs.
Test religiously, but test the right things
Caples spent 50 years testing headlines. He had a framework for how to test that is still valid:
Only test one variable at a time. Same body copy, different subject line. Same everything except the one thing you’re measuring. Any other approach produces ambiguous results.
Use large enough samples. Caples wouldn’t draw conclusions from small tests. He ran ads in multiple publications to large audiences before declaring a winner. In cold email, you need at least 200 sends per variant, and ideally more.
Measure what matters. Caples measured response rates (the equivalent of reply rates), not readership or “awareness.” He cared about the outcome, not proxies for the outcome. In cold email, open rate is a proxy. Reply rate is the outcome. Don’t optimize subject lines purely for open rate if the opens aren’t converting to replies.
Run the test to completion. The temptation is to call a winner after the first promising results. Caples resisted this. Early leaders don’t always win. Let the test run to statistical significance.
Archive every result. Caples kept detailed records of every test. These records became the institutional knowledge of what worked for this audience in this category. Over time, you build a library of winning patterns that new copy can be written against.
One of the most valuable things an outbound team can own is a log of every subject line they’ve tested, with the open and reply rates attached. Most teams don’t maintain this. The result is that every new campaign starts from scratch, repeating experiments that have already been run, relearning lessons that were already learned.
The best subject lines from Caples’s principles, applied
Bringing together self-interest, news, and curiosity, here are the patterns that Caples’s principles would generate for a cold email program:
Self-interest patterns:
book more meetings from the same listoutbound that doesn't require more SDRsfix your cold email reply ratescold email infrastructure that doesn't require babysitting
News patterns:
saw [Company] posted three SDR rolesyour deliverability after 60 days of sendingwhat changed with Gmail filtering in January
Curiosity plus self-interest patterns:
why your reply rates drop after month twothe part of cold email setup most teams skipwhat we found when we audited 50 cold email programsone thing your competitors are doing differently on outbound
How I / How to patterns:
how to send cold email that doesn't land in spamhow we got a 9% reply rate on a 500-person listhow to tell if your domain reputation is declining
Notice what’s not on this list: anything cute, anything that tries to be clever, anything that obscures what the email is actually about. Caples’s testing made him deeply skeptical of cleverness. He’d seen it fail too many times.
The subject line test you should run this week
Take your current top-performing sequence. Identify the subject line on email one.
Write three alternatives:
- One pure self-interest version (directly promises the reader a benefit)
- One curiosity-plus-self-interest version (creates a gap the reader wants to close, connected to something they care about)
- One “how I” or “how to” version (promises useful information without requiring belief)
Split your next 800 sends evenly across all four variants (current plus three new). Run to completion. Keep the winner. Retire the others.
Do this every month. After six months, you’ll have a subject line grounded in evidence rather than intuition, and a log of 18 tests that tells you exactly what this audience responds to.
That’s the Caples method. He did it for 50 years. It works.
The bottom line
Caples’s central lesson was deceptively simple: stop guessing and start testing. Stop writing what you think sounds good and start finding out what actually gets responses.
Every subject line is a hypothesis. The inbox is the lab. The open rate and reply rate are the data.
Write headlines that promise something the reader wants. Make them specific. Test them against each other with enough volume to trust the result. Keep the winners. Throw away the losers.
After 50 years of tests, Caples had strong opinions about what worked. After enough tests of your own, you will too.