"Cold Email Benchmarks 2026"

Healthy B2B cold email in 2026 looks like this: a 1–5% reply rate, 15–30% of replies positive, 50–65% of positives booking, a 60–85% show rate, bounces under 2% — and roughly 1–6 held meetings per 1,000 well-targeted contacts worked end to end, about 2.8 when every link lands mid-range. Most published benchmarks are vendor marketing or averages dragged down by spray-and-pray senders; these are the working ranges we hold our own campaigns to.

More useful than the ranges themselves is what each number diagnoses when it's off.

First, Say What the Number Is a Share Of

A reply rate means nothing until you say what it is a percentage of, and this is where most published benchmarks quietly fall apart. The same campaign can honestly report a 4% reply rate or a 2% reply rate depending on whether the denominator is contacts or emails sent.

Every rate in this article is expressed per contact, across a full sequence. That is the only definition under which the numbers multiply together, and multiplying them is the whole point of having them. If your sequencer reports against emails sent, expect it to show roughly half these figures on a two-email sequence, because each contact gets counted twice in the denominator and a contact who replies stops receiving the rest.

What Reply Rate Should You Expect?

1–5% of contacts emailed. Under 1% after a thousand contacts means the list, the offer, or the copy is wrong — usually the list. 3%+ on a cold list is genuinely good. Above 8%, check your counting: you're probably including bounces, auto-replies, and out-of-office messages, which aren't replies.

How Many Replies Should Be Positive?

15–30% of replies. Of the humans who answer, how many say some version of "tell me more"? Below 15%, your targeting is off — you're reaching people, just the wrong ones. This ratio is the fastest signal for whether an ICP hypothesis is right.

How Many Positive Replies Become Meetings?

50–65% of positive replies book, when they're answered the same day. A positive reply is someone expressing interest or asking a question about your offer. Most of those should reach a calendar, and if fewer than half do, the problem is almost always the speed of the follow-up rather than the targeting. Positive replies decay by the hour.

Be careful with the much lower figures often quoted for this step. Numbers in the 10–20% range describe what share of all replies book, which is a different and much larger denominator, or they describe campaigns where interested people waited days for an answer. Applying an all-replies figure to positive replies alone understates a healthy funnel by a factor of three or more.

What Show Rate Is Normal?

60–85% of booked meetings hold. B2B demos average 30–50% no-shows without reminders. With a confirmation, a 24-hour reminder, a 1-hour reminder, and an SMS, top performers hold no-shows under 15–20%. If your vendor bills on booked rather than held meetings, notice that they're paid either way — the difference between those models matters.

What Do the Ranges Multiply Out To?

This is the test a set of benchmarks either passes or fails: chain them together and see whether the answer matches the headline figure. Ours does.

The arithmetic extremes are wider than the working range: every link at its worst gives about 0.5 held meetings per 1,000 contacts, and every link at its best gives about 8. Neither is worth planning around, because it takes all four numbers landing at the same end at once. Plan on the midpoint, which is 2.8 held meetings per 1,000 contacts, and treat anything above 6 as a claim to check rather than a target to expect.

What that is worth in revenue depends on your contract value and your close rate, which is the arithmetic in the outbound revenue calculator.

What Bounce Rate Is Safe?

Under 2%, ideally under 1%. The deliverability tripwire. Past 2–3%, inbox providers start treating the whole domain as suspect, and every future send pays for it. This number should be watched per-send, not reviewed monthly.

Which Metrics Are Vanity?

Open rate died as a reliable metric when Apple Mail privacy protection started auto-loading pixels — and tracking pixels themselves hurt deliverability. Emails sent is an input, not an outcome; celebrate it and you optimize toward the spam folder. "Leads generated" means nothing until it's defined — a scraped contact, an opener, and a booked meeting have all been sold under that word. (What each agency model actually charges for those words is in how much does a cold email agency cost.)

How Should You Actually Use Benchmarks?

As diagnostics, not scoreboards. The point isn't grading a campaign after the fact — it's knowing which part is broken while there's still budget to fix it. Low reply rate → list problem. Good replies, few positives → targeting problem. Positives that never book → speed problem. Booked but not held → reminder problem. Each has a different fix, and funnel-wide averages hide all four.

That's also the standard to hold a vendor to: weekly reporting, per-segment, including what isn't working. An agency that only reports good weeks is reporting a story. The other five questions worth asking are in how to vet an outbound agency.

Correction, 9 September 2026

This article previously gave 10–20% as the share of positive replies that book, and separately claimed 3–6 held meetings per 1,000 contacts. Those two figures cannot both be right: the first chains out to well under one meeting per 1,000. The booking figure was the error. It has been corrected to 50–65%, along with the headline range, which now matches what the chain actually multiplies out to. We publish these numbers so they can be checked, and that only means something if we fix them when the check fails.