Last month we sent 13,898 emails across 11 campaigns. We got 164 total replies and 80 positive responses.
That is a 1.18% reply rate and a 0.5% positive reply rate, with zero personalisation and zero AI.
Talk to most cold email experts and that is a failure. Shut it down, fix the copy, buy a personalisation tool. Here is my honest reaction to those numbers: who cares?
Not as a dodge. I mean it as a question about what the number is for.
Why the benchmark is the wrong question
Every page you will find on this query is a table of averages that makes you feel bad about your number, followed by the thing that will fix it, sold by the people who published the table.
The table isn't useless. It is just answering a question that does not decide anything. If what you actually want to know is whether your email is any good, the grader answers that directly and in about thirty seconds.
Cold email is the cheapest channel to scale. That is the whole economic point of it. If you have a large addressable market, and most B2B companies do, the maths works at almost any reply rate as long as the pipeline value is there.
So: who cares whether you send 100 emails to get the pipeline or 10,000 emails to get the pipeline? The sends are close to free. The pipeline is what you were buying.
A 1.18% reply rate on 13,898 emails produced 80 positive conversations, real pipeline across several ideal customer profiles, and live data on which segments respond to which message. Measured as a reply rate it looks poor. Measured as market feedback at scale it was one of the more useful months we have had.
The number I actually track
Sends per positive reply.
My best week was 738 sends per positive reply. My worst was 6,355. Same product, same person writing the copy. The only variable was how tight the targeting was.
That is the metric worth putting on a dashboard, for two reasons.
It is an input you control. Reply rate is an outcome that moves when the market moves, when a provider changes something, when a competitor emails your list the week before you do. Sends per positive reply tracks the thing you actually change, which is the list.
And it prices the work. If it takes 6,355 sends to get one positive reply, you can decide whether that is worth it. A percentage does not let you make that judgement because it hides the volume.
Why the economics work at a low reply rate
It is worth being precise about the argument, because "who cares" on its own sounds like an excuse.
Cold email is the cheapest channel you have for reaching people who have never heard of you. Not the most effective per contact. The cheapest per contact, by a distance, and those are different claims.
That changes what a reply rate means. In a channel where each attempt is expensive, a low conversion rate is fatal, because you cannot afford the volume that makes it work. In a channel where each attempt costs almost nothing, the conversion rate is one input among several, and the one that actually decides things is whether the pipeline at the end is worth more than the effort at the start.
So if you have a large addressable market, and most B2B companies do, the maths holds at a reply rate that would be a disaster anywhere else.
The second thing you are buying is less obvious and I value it more. That month of 13,898 sends was not only pipeline. It was messaging tested across several ideal customer profiles at once, live, with real buyers, in a fortnight. It told us which segments were in market now and which weren't, which is a question most companies answer with a survey or a guess.
That is the argument for sending at volume, and notice it is not "send more to get more replies." The same platform data showed tripling volume producing fewer replies. It is that a cheap channel run against a well defined list buys you market feedback at a speed nothing else matches. The reply rate is the price of the information, not the point of it.
The range is enormous and nobody publishes the ends
Benchmark tables flatten the range, which is the most misleading thing about them. Here are both ends from my own campaigns in the same year.
One week: 5 meetings from 150 sends in six hours, at an 83% positive reply rate.
Another month: 13,898 sends, 0.5% positive.
Another week: one meeting from 13 sends.
Those are not different skill levels. They are different lists, different offers and different tiers of account. If your reply rate is 1%, you are not failing at a task whose correct answer is 5%. You are running a different campaign shape from the one in the table, and the table cannot tell you which.
What the volume data actually says
We looked at 8 million cold emails: 5.25 million prospects, across 20 businesses, over 8 weeks. Real platform data, not a survey.
Three things from it that change how you should read your own numbers.
Tripling send volume produced fewer total replies, not more. If your instinct when replies drop is to send more, that instinct is measurably wrong. More sends is a lazy substitute for better targeting.
Azure and Google mailboxes pull nearly twice the human reply rate of Outlook. A chunk of what people read as a copy problem is a provider choice. The setup maths is in how many inboxes for cold email.
The first email plus three follow ups captures almost everything. After that you are not following up, you are annoying people. If you are running an eight step sequence, steps five to eight are not improving the reply rate in that table, they are costing you reputation.
When the reply rate does matter
One case, and it is diagnostic rather than comparative.
If reply rate drops significantly on a campaign where the copy and the list have not changed, something broke. That is worth acting on immediately. It is a change detector, not a scorecard, and the order to work through is in cold email not getting replies.
Comparing your 1.4% to someone's published 5% tells you nothing, because you do not know their list, their tier, their offer or whether the number is real. Comparing this week's 1.4% to last week's 3% on your own campaign tells you something immediately.
The numbers that actually matter
Meetings, and what they turn into.
For context on what this looks like when it works: over the stretch I wrote up as a guide, we went from 4 meetings a week to 38. These days we book around 30 cold meetings a week on autopilot.
And on the other side of it, from people who bought the system rather than built it. Jon Rudoe, Chief Commercial Officer at Nous, describes the engine delivering 20 meetings a week on autopilot, which removed the need to hire SDRs so they hired AEs instead. Dom Bowcock at Bowimi puts it at an extra £200K in ARR in the first six months.
None of those are reply rates. That is the point.
So what is a good cold email reply rate?
The honest answer is that it is the wrong question, and the useful version is this: good is whenever the pipeline you create is worth more than the sends and the time it took to create it.
Track sends per positive reply. Watch it move when you tighten the list. Use reply rate as an alarm rather than a grade. And stop reading the tables.