Search for a good cold email reply rate and you’ll find an average. Three percent, five percent, eight and a half percent, depending on which vendor published it. Each is probably real, and each is a recruiting agency, a software company and a wholesaler averaged together.
The number that matters is the one for your audience, and nobody can give you that in advance. What I can give you is every number I have, with where it came from, so you can find the row that looks most like you.
Numbers from campaigns I’ve run
For a B2B campaign to companies with repeat customers, I tell people to expect 1 to 1.5% of those emailed to reply, and around 20% of those replies to be interested. Ten to fifteen replies in a 30 to 45 day pilot is a good start.
Some campaigns do better. The best I’ve run went to 15,700 sends with a 4.92% reply rate and 178 opportunities. A set of four campaigns to one audience, each to 900 to 1,900 contacts, brought back 35, 40, 10 and 37 replies. One small campaign got 13 interested replies from 60 emails, which is the best rate I’ve seen and also a list of 60 people, so I don’t quote it as typical.
One did a lot worse. 1,300 emails to consumers returned nothing. Not a single reply. That was enough for me to stop sending to consumers.
What people who send for a living report
These are figures individual practitioners have given for their own work. Not studies.
| Audience | Figure given | Who |
|---|---|---|
| Recruiting and staffing outbound | 1% positive replies is a strong result | An agency owner who runs staffing campaigns |
| Pitching investors | 20% positive replies | A founder raising money |
| Narrow niche, small list | 2.5 to 5% total replies, consistently ahead of broad targeting | A solo operator on his own campaigns |
| Hand researched, under two hundred contacts | 15 to 25% replies | An agency owner |
| Volume, database list, tens of thousands | 1 to 2% replies | Same agency owner |
| B2B with repeat customers | 1 to 1.5% replies, about 20% of those interested | My own campaigns |
The 15 to 25% figure travels a long way online. The person who gave it was describing lists of a hundred or two hundred companies, each one researched by hand. He didn’t claim it held at any larger volume, and I haven’t seen anyone show that it does.
Positive replies, not replies
Most benchmarks count every reply. “Take me off your list” is a reply. So is “wrong person.” The number that pays is positive replies, meaning someone asked a question or agreed to talk.
I set one number before a pilot starts: under 0.5% positive replies and the audience was wrong, so we stop rather than buy a second month. Above it, the audience works and the emails get better from there. That line gets agreed before anything is sent, so nobody argues about it in week five. Your line may be different. Have one.
How big a sample before you believe it
Two agency owners gave me rules of thumb that point the same way. One sends each angle to 800 to 1,000 contacts before deciding whether it works. The other treats a first campaign as 1,000 to 2,000 leads before he can tell anything, and aims for one positive reply per 300 to 400 contacts when there’s no prior data.
At the rates above, 200 emails at 1% is two replies. Zero is within normal variation. So “we sent 200 emails and got nothing” doesn’t tell you much either way.
Volume or precision
There’s a long running argument about whether to send a lot to a broad list or a little to a narrow one. The table suggests both work, at different rates.
One person in that debate argues the question itself is wrong. The lever he sees is where the list came from. A list bought from the same database as every competitor performs like a database list whatever you do to the copy. A list from somewhere nobody else has looked performs closer to a hand researched one. I don’t have the data to prove that, and it matches how I think about building a list.
Questions we get
What’s a good open rate? I don’t measure it. Open tracking means a tracking pixel in every email, which hurts deliverability, and a large share of recorded opens are security scanners rather than people. Replies are the number to run on.
Is 2% good? For B2B to a database list, yes. For a hand researched list of 150, it’s low. Compare against the row closest to you, never the overall average.
How long until I know? On my campaigns, first replies usually arrive in the first week of sending. A verdict on the audience takes the full 30 to 45 days and around a thousand contacts.
We got zero. Is cold email dead? Usually something else broke first. Zero replies of any kind, including negative ones, is the pattern that goes with emails that never landed. Check that before you conclude anything about the channel.
