Claude Hopkins Invented Scientific Testing in 1923. Cold Email Still Hasn't Caught Up.
In 1923, Claude Hopkins wrote a book called Scientific Advertising. It was 234 pages. David Ogilvy said every person in advertising should read it seven times. Gary Halbert kept a copy at his desk for decades.
The central argument of the book: advertising should be measured, tested, and optimized the same way a scientist runs an experiment. Gut instinct is a starting point. Data is what you run on.
Hopkins arrived at this through his career in mail-order advertising, where every piece of copy had a coupon that could be tracked. He knew which ads worked and which didn’t because he could count the coupons. He ran the first split tests. He pioneered the use of free samples. He invented the idea of “reason-why” copy. He was the first person to prove, with numbers, that one approach to advertising outperformed another.
Cold email is the closest modern equivalent to the mail-order world Hopkins worked in. Every email has a trackable outcome. Every variable can be isolated. Every sequence can be improved based on data.
And yet most cold email programs run on the same copy indefinitely, changing things randomly when results decline, with no systematic approach to improvement.
Hopkins would find this inexcusable.
Scientific Advertising: what it actually means
Hopkins’s scientific method was not complicated. It was disciplined.
The process:
- Run an ad with a specific, trackable offer
- Count the responses
- Change one variable
- Count the responses again
- Keep the winner
- Repeat
What made it revolutionary was not the method itself but the commitment to evidence over opinion. Hopkins worked with clients who had strong opinions about what their ads should say. He didn’t argue with opinions. He tested them.
“Almost any question can be answered cheaply, quickly, and finally by a test campaign. And that’s the way to answer them. Not by arguments around a table.”
In cold email, this means:
Every meaningful variable can be tested. Subject lines, opening hooks, proof points, CTAs, sequence length, follow-up timing, email length. You don’t have to guess which one is better. You can find out.
What you cannot do is change multiple variables at the same time and then draw conclusions. If you rewrite the entire email and open rates go up, you don’t know which change drove the improvement. One variable. One test. One winner.
This is the Hopkins discipline that most cold email teams don’t have. They’ll rewrite a sequence when it stops working, changing everything at once, and then start again from scratch the next time results decline. There’s no cumulative learning. The program doesn’t get better over time because there’s no systematic way to build on what worked.
Reason-why copy
This is Hopkins’s most important contribution to copywriting, and the most directly applicable to cold email.
The principle: whenever you make a claim, give the reason for it. Not just “our deliverability is excellent” but why it’s excellent. Not just “book a call” but why you should book a call right now.
Hopkins discovered this because he tested it. Reason-why copy consistently outperformed copy that just made claims. Giving the reader a logical explanation for why something is true makes the claim believable. And believability is what drives the response.
The mechanism is simple: people are skeptical of claims, especially from strangers. A claim without a reason sounds like every other ad they’ve ever ignored. A claim with a specific reason sounds like something a knowledgeable person told them.
Without reason-why:
“Our cold email program gets exceptional results.”
With reason-why:
“Our reply rates are consistently higher than industry average because we build separate infrastructure for every client. No shared IP pools. No shared domains. Your sending reputation doesn’t depend on anyone else’s behavior.”
The second version is the same claim, but now the prospect understands why it might be true. They can evaluate the logic. If the reason makes sense to them, the claim becomes credible.
In cold email, reason-why copy shows up in four places:
Why you’re emailing this specific person now:
“Reaching out because [Company] just posted two SDR manager roles. That usually signals either the function is scaling or it’s struggling. Either way, it’s the moment when infrastructure becomes urgent.”
Why your approach works:
“We run cold email on isolated infrastructure per client. No shared pools. That’s why our clients maintain good deliverability when agencies using shared infrastructure are getting filtered to spam.”
Why the numbers are what they are:
“We booked 41 meetings in 90 days for a Series A security company. The reason: we spent the first two weeks on list quality and signal identification before writing a word of copy. Most programs skip that.”
Why they should respond now:
“We typically have a two to three week onboarding queue. If you’re looking to scale outbound in Q2, now is when the conversation makes sense.”
Each reason-why turns a claim into an argument. Arguments are harder to dismiss than claims.
The specificity principle
Hopkins proved through testing that specific numbers outperformed round numbers. “57 varieties” outperformed “many varieties.” “Loses 5 pounds in 7 days” outperformed “rapid weight loss.” Specific claims performed consistently better than vague ones across hundreds of tests.
The reason: specificity signals that the claim is real. Vague claims can be made up. Specific ones suggest someone actually measured something.
Cold email is full of vague claims:
- “We help companies increase their pipeline”
- “Our clients see significant improvements in reply rates”
- “We’ve worked with many companies like yours”
- “Proven results in your industry”
These are signals to the reader that the sender doesn’t have real evidence. They’re reaching for adjectives because they don’t have numbers.
Replace every vague claim with the specific version:
| Vague | Specific |
|---|---|
| ”significant pipeline increase" | "43 qualified meetings in one quarter" |
| "improved reply rates" | "went from 1.2% to 6.8% reply rate in 60 days" |
| "many companies like yours" | "14 B2B SaaS companies between 50 and 200 employees" |
| "proven results" | "8 of our last 10 clients renewed after month three” |
The specific version is harder to write because you have to actually know the numbers. That’s the point. If you don’t know your numbers, you don’t have a case yet. Go find the numbers first.
One important nuance Hopkins understood: you don’t round up. “57 varieties” sounds specific and accurate. “60 varieties” sounds like a round number someone made up. Specificity requires precision, not impressiveness.
Free samples and the trial offer
Hopkins pioneered the use of free samples in mail-order advertising. His insight: the best way to get someone to buy a product is to get them to use it. Once they’ve experienced it, the sale is already made.
He used samples for everything. Shampoo. Toothpaste. Canned food. Each sample had tracking on it so he knew which samples converted to sales and which didn’t. The sample was both a sales tool and a research instrument.
In cold email, the equivalent is the low-risk offer.
Most cold email asks for a lot: a 30-minute demo, a full sales call, a meeting with your entire team. For a prospect who doesn’t know you and hasn’t asked for your help, this is a big ask. They have to commit time, prepare, and mentally rehearse for what’s probably going to be an awkward first conversation.
The Hopkins approach: what’s the smallest possible thing you can offer that lets them experience value before committing?
Some options that work:
The audit or teardown: “Happy to give you a 10-minute teardown of your current cold email setup. No commitment. Just feedback.” This is a sample of your expertise. If the feedback is good, the sale is already half done.
The specific piece of relevant data: “Put together a quick overview of the deliverability setup issues we’re seeing at companies your size. Want me to send it over?” A piece of genuinely useful information costs them nothing to receive and starts the relationship.
The 15-minute no-pitch call: “Not trying to pitch you on a long engagement. If you’re 15 minutes, happy to share what we’re seeing in outbound programs at your stage and you can decide if it’s worth a follow-up.” The explicit “not pitching” framing lowers the guard.
Each of these is a sample. Hopkins would track which type of sample converted to a booked call, and optimize from there.
Track your coupons
Hopkins’s ability to improve his campaigns was entirely dependent on tracking. Every ad had a coupon. Every coupon was coded. He knew which publication, which headline, and which offer produced each response.
In cold email, you have better tracking than Hopkins ever had. Open rates, click rates, reply rates, positive reply rates, meeting booking rates. Every variable in the sequence can be measured.
But measurement only matters if you use it to make decisions.
What Hopkins would insist on:
Track by variable, not just by campaign. Knowing that a campaign got 4% reply rate doesn’t tell you anything useful. Knowing that subject line variant A got 8% reply rate and subject line variant B got 2% tells you something you can build on.
Track at each stage of the funnel. Open rate tells you about the subject line and sender. Reply rate relative to open rate tells you about the copy. Meeting booking rate relative to reply rate tells you about the offer and the ask. Each stage has its own variables.
Set a sample size before you call a winner. Hopkins was firm on this. You don’t declare a winner at 20 sends. You set a minimum threshold before you start the test (typically 200 sends per variant for cold email) and you don’t make decisions until you hit it.
Keep a log. Every test, every result, every winning variant. This is your institutional knowledge. When a sequence starts underperforming six months from now, you’ll be able to trace what changed instead of starting from zero.
The compounding advantage of scientific testing is real. A program that systematically improves by 10% per month through disciplined testing is a fundamentally different thing from a program that gets relaunched every quarter because nobody knows why it stopped working.
The Hopkins cold email audit
Here are the questions Hopkins would ask about your current program:
On testing:
- Have you run a split test in the last 30 days?
- Do you test one variable at a time, or do you rewrite everything when things aren’t working?
- Do you have a log of what you’ve tested and what won?
On specificity:
- Can you name the specific result you’ve produced for a client in one sentence with real numbers?
- Is every claim in your email backed by a reason-why?
- Have you used any vague adjectives that a competitor could use to describe their product?
On the offer:
- Does your CTA ask for more than a prospect is likely to give on a cold introduction?
- Have you offered any form of sample, audit, or low-risk preview of your value?
On tracking:
- Do you know your reply rate broken down by email in the sequence?
- Do you know which subject line variant outperforms the others?
- Do you know what your positive reply rate is (interested replies vs. unsubscribes)?
If you can’t answer most of these, your program is running on opinion rather than evidence. Hopkins would say you’re not doing advertising. You’re doing guesswork with a logo on it.
The bottom line
Hopkins’s core belief was that advertising was a science that could be learned and improved through systematic measurement. He had no patience for intuition that couldn’t be tested, opinions that couldn’t be verified, or copy that couldn’t be tracked.
Cold email is the most measurable outbound channel that exists. Every variable is testable. Every outcome is countable. Every sequence can improve.
The gap between a cold email program that plateaus and one that compounds is usually not budget or creative talent. It’s the discipline to test one thing at a time, measure what happens, and build on what works.
Hopkins figured this out a hundred years ago. The tools are better now. The discipline is still the same.