Cold email copywriting is not persuasion writing. It is a compression problem: you have a subject line, roughly 60–80 characters of preview text, and about eight seconds to earn the rest. Almost every rule that follows exists because of that constraint, and almost every bad cold email breaks it by trying to be complete instead of being specific.
This guide covers the four parts of the email, the structures that work at each, and the failure patterns that show up in nearly every audit we run.
The Only Job Cold Email Copywriting Has
All cold email copywriting serves one goal, and it is not the obvious one. Not to sell. Not to explain your product. To make a reply feel like less work than deleting. That reframing changes most drafting decisions on its own: it makes you cut the second paragraph, drop the calendar link, and replace “book a 30-minute discovery call” with a question someone can answer in one line.
Under 100 words, one value proposition, one ask. Everything you cut belongs later in the sequence, and most of it belongs nowhere.
Part 1: The Subject Line
The job of email one is not to sell. It is to make replying feel like less work than deleting.
The subject line’s job is to be opened, not to be clever, and the two are frequently in tension. What works is short, lowercase or sentence case, and specific enough to be about them rather than about you. Three to five words is the working range.
What reliably fails: anything that reads like a marketing campaign, all-caps, stacked punctuation, and “Quick question” — which has been used so heavily that some filters now treat it as a signal in itself. Our subject line guide covers the tested patterns.
Part 2: The Opening Line
This is where most cold email dies, and it dies in a predictable way: the first line is about the sender. “I’m reaching out because”, “I came across your profile”, “Hope this finds you well” — all of them spend the most valuable sentence in the email establishing nothing.
The opening line should demonstrate a specific piece of knowledge that took effort to acquire, in a way that is obviously not a mail merge. A trigger event, a technology they run, a number from their world. The test is simple: if the sentence would be true of a hundred other companies, it is not an opening line, it is filler. Our breakdown of hooks and opening lines covers the patterns, and the eight-second rule of inbox scanning explains why the window is so unforgiving.
Part 3: The Value Proposition
One sentence, in their vocabulary, about an outcome rather than a capability. The most common failure is describing what the product does instead of what changes for them — and the two sound identical to the person who built the product and completely different to the person reading the email.
There is a real strategic choice here between leading with the pain and leading with the outcome. Pain-led versus value-led covers when each wins; the short version is that pain-led works when the problem is already conscious, and value-led works when it is not.
Part 4: The Call to Action
| CTA type | Example | Effort to reply | Use when |
|---|---|---|---|
| Interest check | “Worth a look?” | 1 word | Default for email one |
| Specific question | “Are you still handling this in-house?” | 1 line | You have a real hypothesis |
| Resource offer | “Want the teardown?” | 1 word | Later in the sequence |
| Meeting request | “15 minutes Thursday?” | Calendar check | After a positive signal |
| Calendar link | Booking URL in email one | High | Effectively never |
The ranking is by reply friction, and reply friction is the variable you control most directly. A calendar link in a first email asks a stranger to commit time before they have agreed there is a topic. Our CTA guide covers the progression across a sequence.
Personalisation That Scales
This is where cold email copywriting has changed most. Token personalisation — first name, company name, job title — no longer reads as personal, because everyone has it. What still works is segment-level specificity: find a cluster of accounts sharing a real attribute and write one genuinely informed email for that cluster. Fifty emails that are right about a shared situation beat a thousand that are technically correct about a merge field. Personalisation at scale covers the mechanics.
The Cold Email Copywriting Failure Patterns
subject line 3-5 words about them, not you
opening line 1 sentence specific enough to be unmailmergeable
value proposition 1 sentence an outcome, in their vocabulary
call to action 1 question answerable in one word
----------
total < 100 words ← everything else goes later
- Writing to the buyer you wish you had. Copy addressed to someone already convinced there is a problem, sent to someone who is not.
- Front-loading credibility. Logos and funding announcements in email one. Nobody is evaluating your credibility yet; they are deciding whether this is about them.
- Two asks. “Let me know if useful, or grab time here” halves the reply rate by making the reader choose.
- Follow-ups that just bump. The second email should be a different angle. “Just floating this back up” adds volume and no information.
- Length creeping back. Every round of internal review adds a clause. Re-cut to under 100 words before sending.
How to Actually Improve the Copy
Test one variable at a time, minimum 100 sends per variant, judged on positive reply rate rather than opens — privacy features have made open rate close to meaningless. Lock the hook first, then the value proposition, then the CTA, because the hook gates everything downstream and a better CTA on an unread email changes nothing.
Compliance sits underneath all of it: the FTC's CAN-SPAM guidance sets the US floor, GDPR governs EU contacts, and neither is optional because your copy is good.
Further Reading
Handling objections in replies
Cold email for founders and CEOs
Reply rate benchmarks by industry
GDPR, CAN-SPAM and CASL compliance
The Bottom Line
Good cold email copywriting is mostly subtraction. Four parts — subject, opening line, value proposition, CTA — each doing one job, totalling under 100 words, aimed at making a reply cheaper than a delete. The specificity of the opening line and the friction of the ask explain most of the variance between a 1% reply rate and a 10% one.
The trap is treating copy as the lever when it is the last one. If placement is broken or the list is loose, better writing produces a better-written email that nobody relevant reads. Fix those first, then this guide is worth working through properly. If you would rather have the whole system built and tested for you, book a meeting.
FAQ
How long should a cold email be?
Under 100 words for the first, under 80 for follow-ups. Length correlates negatively with reply rate in essentially every test we run, and needing 300 words to explain your value usually points at a positioning problem rather than a writing one.
Should I include a calendar link in the first email?
No. It asks a stranger to commit time before agreeing there is a topic worth discussing. Ask an interest-check question first and send the link once someone has replied positively — the link converts far better at that point anyway.
Does AI-written cold email still work?
As a drafting aid, yes. As the finished article, it is now recognisable, because recipients have seen enough of it to pattern-match. The useful division is AI for structure and volume, a human for the specific observation that makes the email worth reading.
How many variants should I test at once?
Three to five hooks, one variable at a time, at least 100 sends each. Fewer than 100 and the result is noise; more than five variants and you are splitting the sample too thin to reach significance inside a campaign cycle.
