Cold email benchmarks circulate widely and are mostly unreliable, because they come from vendors with an interest in the numbers looking good. What follows is a conservative picture of what a competent, well-targeted campaign achieves.
Realistic numbers
Open rate: 30–50% for a well-targeted campaign with clean deliverability. This figure has also become less meaningful as privacy features inflate opens by pre-fetching images, so treat it as a rough deliverability indicator rather than an engagement measure.
Reply rate: 3–8% for a good campaign. Above 10% usually means a very narrow, very well-researched list, or an existing relationship.
Positive reply rate: 1–3%. Replies that express interest rather than declining.
Meeting rate: 0.5–2% of emails sent.
The practical implication: to book 20 meetings you need roughly 1,500–2,000 well-targeted emails, not 300. Campaigns fail most often because the volume required was underestimated and the effort stopped after two weeks.
What moves reply rate, in order
1. List quality. By a wide margin the largest factor. The right person at the right company with the problem you solve will reply to a mediocre email. The wrong person will not reply to a brilliant one. Most effort spent on copy would return more if spent on targeting.
2. Relevance of the opening. The first line decides whether the rest is read. It must demonstrate you know something specific about them — not “I saw you’re in software”.
3. Clarity of the ask. A specific, small request outperforms a vague or large one. “Worth a 15-minute call?” beats “let me know if you’d like to learn more”.
4. Length. Under 125 words. Cold emails are read on phones between other things.
5. Follow-up. A sequence of three to four messages roughly doubles total replies compared with a single send.
6. Subject line. Matters less than most people think once deliverability is sound. Short, plain and specific beats clever.
Structuring the message
Four parts, in order.
The relevance line. Why them, specifically. A trigger event, something they published, a role they are hiring for, a problem their situation implies.
The problem. One sentence naming something they plausibly experience. Not a product description.
Brief evidence. One specific result from a comparable company, with a number.
The ask. One question, easy to answer, low commitment.
What to leave out: your company’s founding story, a feature list, more than one link, an attachment, and anything that reads as though it was merged from a template.
The sequence
Three to four messages over two to three weeks.
The first is the full message described above. The second, three or four days later, adds a different angle or a piece of evidence — not “just bumping this up”. The third, a week later, is short and offers a different, smaller ask. The fourth closes the loop politely and says you will stop.
That last message consistently generates replies, partly because it is the only one that asks nothing. Stopping after it is important: continuing past four messages generates complaints, which damages deliverability for everything else you send.
Deliverability for outbound
Cold email has a higher risk profile than marketing email, and the technical setup matters more.
- Use a separate domain. A lookalike domain for outbound, so a reputation problem cannot affect your main domain’s mail.
- Authenticate it fully. SPF, DKIM and DMARC on the sending domain. See email deliverability.
- Warm the domain over several weeks before sending volume.
- Keep per-mailbox volume low. Commonly 30–50 a day per mailbox, spread across several mailboxes rather than one sending hundreds.
- Verify addresses before sending. A bounce rate above a few percent damages reputation quickly.
- Send plain text. No images, no tracking pixels, minimal links. It looks like a personal email because it is supposed to be one.
The legal position
Rules differ substantially by jurisdiction and change, so this is orientation rather than advice. In broad terms, several regimes permit unsolicited business-to-business email with conditions — accurate sender information, a working opt-out, and a clear indication of the purpose. Others, particularly where consent-based rules apply to all recipients, are considerably stricter.
The practical minimum regardless of jurisdiction: identify yourself and your company accurately, provide a genuine way to opt out, honour it immediately, and do not send to people who have asked you to stop. Check the specific requirements for the countries you are mailing into before scaling a programme.
Measuring it properly
Track the whole funnel rather than open rate: emails sent, delivered, replied, positively replied, meetings booked, opportunities created, revenue closed.
The ratio that matters is emails sent per opportunity created, because that is what tells you whether the channel is economic once you count the time. Many outbound programmes look healthy at the reply stage and fall apart at the opportunity stage, which usually means the targeting is attracting polite interest from people who will never buy.
Review by segment. A programme averaging 4% reply may be 9% in one segment and 1% in another, which is an instruction to narrow the list rather than rewrite the email.
Building the list
Since list quality dominates everything else, it deserves most of the preparation time.
Start from the ideal customer profile and build the company list first, before thinking about contacts. Then identify the right role within each — usually the person who owns the problem, not the most senior person available. Senior contacts forward or ignore; the owner of the problem replies.
Layer a trigger where you can. Companies hiring for a relevant role, announcing a funding round, opening a new market or publicly discussing the problem are measurably more responsive than the same companies at a random moment.
Then verify the addresses and remove anything uncertain. A list of 300 verified, well-chosen contacts outperforms 3,000 scraped ones on every measure including total meetings booked, and it protects the sending domain.
When outbound is the wrong channel
Three situations where a cold email programme will disappoint regardless of execution.
Low contract values. The economics do not work. If a deal is worth a few hundred pounds a year, the hours required per meeting exceed the margin.
A product that needs explaining. If the value cannot be conveyed in two sentences to someone who has never heard of the category, cold email is the wrong first touch. Content and demand creation have to come first.
A very small market. With fifty possible customers, email is too blunt. Those relationships are built individually, through introductions and events.
In each case the alternative is not no outbound, but outbound that starts from a warmer position — a referral, a community, an event, or content the prospect has already read.
Reviewing a campaign that is not working
When reply rates are poor, the diagnosis order matters because the fixes are very different.
Check deliverability first. If open rates are below 20%, the messages are not arriving and nothing about the copy is the problem.
Then check the list. Pull twenty contacts at random and ask honestly whether each is the right person at a company with the problem. If more than a few are not, the list is the issue.
Only then look at the copy, and test the opening line before anything else, since it determines whether the rest is read. Testing subject lines while the list is wrong is the most common wasted effort in outbound.
Frequently asked questions
What is a good cold email reply rate?
Three to eight percent for a well-targeted B2B campaign. Figures above ten percent usually indicate a very small, very specific list.
How many follow-ups should a sequence have?
Three to four messages in total. Beyond that, replies fall and complaints rise.
Should cold email be personalised individually?
The opening line should be specific to the recipient. The rest can be templated by segment. Fully individual emails do not scale and fully templated ones do not work.
Is cold email still effective?
It works where targeting is tight and the offer is relevant. Volume-based outbound to broad lists performs steadily worse as filtering improves and buyers receive more of it.
