A cold email gets read in two passes. The first is about a second long and answers one question: is this about me, or is this about them. Almost every email that fails, fails there — and no amount of quality in paragraph three is ever seen.
The first line is the whole decision
Compare two openings that make the same request.
“I hope this finds you well. I’m reaching out because we help companies like yours streamline their revenue operations.”
“You posted a RevOps lead role three weeks ago that lists cleaning up attribution as the first responsibility.”
The second is not better written. It is better grounded. It proves in one line that a human looked at something specific, and it gives the reader a reason to believe the next sentence is also about them.
The first sentence is not an introduction. It is the evidence that this email was written to you rather than at you.
Where the second sentence goes wrong
Having earned a moment of attention with something specific, most emails immediately spend it on a pivot to the sender: “At Acme we help teams like yours…”. The reader has just been told the email was about them and is now being told it was not.
The second sentence should extend the first — the consequence, the question, the thing that follows from what you observed. Your company can wait until sentence four, or until the reply.
| Element | Weak version | Version that works |
|---|---|---|
| Opening | Generic segment description | One dated, checkable observation about them. |
| Relevance | 'Companies like yours often struggle with…' | The specific consequence of the thing you observed. |
| Proof | A customer logo they cannot verify | A concrete claim about what changes, with a number you can defend. |
| Ask | 'Do you have 30 minutes next week?' | A question answerable in one line, or a genuinely smaller step. |
| Sign-off | A postscript with a second offer | Nothing. A second ask dilutes the first. |
Why volume made this worse
Generating a plausible personalised paragraph became free, so everybody started sending them, and readers learned the pattern within about two quarters. A sentence that reads as machine-assembled specificity now does active damage: it signals that no attention was paid, dressed as attention.
The response is not to sound less automated. It is to be more specific — specific in ways that require something real to have been read. A reference to a page on their site, a role they posted, a talk somebody there gave, with enough detail that it could not have been produced without opening it.
The ask should be proportionate
Thirty minutes is a large request from someone who has never heard of you. It is also frequently the wrong request, because a meeting booked out of politeness is worse than no meeting — it consumes an hour and produces a soft no.
A question answerable in one line does more. It is cheap to say yes to, it produces a reply you can build on, and the answer tells you whether a meeting is worth anyone’s time. Ask about the thing you observed: whether the problem you inferred is actually the problem, whether it sits with them.
How to tell whether it is working
- Reply rate, including negative replies — a clear no is a working email meeting the wrong person.
- Replies that engage with your specific observation, which is the only proof the opening landed.
- Meetings that produce a next step, rather than meetings booked.
- How often the observation turns out to be wrong — that is a research problem, not a copy problem.
- Whether the sequence adds anything after the first message. If not, cut it.
Most of the work here happens before anything is written. An opening line that carries a dated, checkable observation exists only if the research produced one, which is why what counts as evidence matters more to reply rates than any copy framework does. In Leads, every company comes back with the sources and dates attached — the first sentence is already written, and you can open the page it came from before you send it.