What Is a Good Cold Email Reply Rate? Benchmarks by Industry and List Size

Ask ten people what the average cold email reply rate is and you'll get ten answers between 1% and 20%. They're all technically true, because none of them are measuring the same thing. One person counts every reply including out-of-office notices, divided by total emails sent across a five-step sequence. Another counts only positive replies divided by leads contacted. Those two numbers can differ by 10x on identical campaigns.

So before benchmarks are useful, you need a definition that doesn't move. Then you need to know which bucket you're in — because a 3% reply rate on a 20,000-lead list and a 3% reply rate on a 200-lead list are two completely different results, and only one of them is a problem worth fixing.

The short answer

For B2B cold outbound to a list you built and verified yourself, measured as unique human replies divided by leads successfully delivered to, across the full sequence rather than a single step, here is the range you should be judging yourself against.

• Under 2% — something is broken. Usually targeting or deliverability, occasionally the offer.
• 2–5% — normal for broad lists (5,000+ leads) with light personalization. Workable but unremarkable.
• 6–10% — good. This is what a tight ICP, a relevant offer, and clean infrastructure produce.
• 11–20% — very good, and almost always on small, hand-built lists under 1,000 leads with real research per account.
• Above 20% — either a warm-ish list (past customers, event attendees, community members), a tiny sample, or you're counting auto-responders as replies.

Positive replies — people who actually want to talk, not "no thanks" or "wrong person, try Dave" — typically run 25–40% of total replies. So a 6% reply rate is roughly a 1.5–2.4% positive rate. From there, about half of positive replies turn into a booked meeting if you respond within a couple of hours instead of a couple of days.

Run the math on a 1,000-lead campaign: 6% reply rate gives you 60 replies, around 20 positive ones, and maybe 10 meetings. That's the honest shape of a decent campaign from a competent sender. If someone promises you 40 meetings from 1,000 cold emails, they are either selling something or counting differently.

Fix the denominator before you compare anything

Three measurement mistakes make published benchmarks useless, and most teams make at least two of them.

Replies divided by emails sent. A four-step sequence to 1,000 leads sends roughly 3,200 emails once people drop out on reply. Fifty replies is 5% of leads but 1.6% of emails. Same campaign, three different-looking numbers depending on how you slice it. Always use delivered leads as the denominator, and always measure at the sequence level. Per-step reply rates are useful for diagnosing which follow-up is dead weight, but they're not the headline number.

Not subtracting bounces. If 12% of your list bounces, those leads never had a chance to reply, and including them in the denominator understates your real performance. Divide by leads that actually received the email. This alone can swing a reported reply rate by half a point or more, and it's the reason verifying every lead before you send changes your reported numbers even when it doesn't change your actual outcomes.

Counting machines as humans. Out-of-office auto-responses, ticketing system acknowledgements, "your message is being reviewed by our security team," and DMARC aggregate reports can easily be 10–20% of everything landing in your reply inbox. If your tool counts all of it, your reply rate looks inflated and your positive-reply ratio looks mysteriously terrible. Filter these out or at minimum classify them separately so you can see both numbers.

The same skepticism applies to open rates, which is why there are no open-rate benchmarks in this post. Since Apple Mail Privacy Protection started pre-loading images and corporate link-scanners started clicking every URL in a message before delivery, opens and clicks sit somewhere between noisy and fictional. Replies and verified site visits are the only engagement signals worth building decisions on.

Benchmarks by list size

List size is the single strongest predictor of reply rate, and it's not because big lists are cursed. It's because relevance and volume trade off directly. You can only research so many accounts per hour, so every additional thousand leads dilutes the average quality of what you're sending.

The practical read: if you need ten meetings a month, 800 well-researched leads will usually beat 8,000 generic ones, and it will do far less damage to your domain reputation along the way. Large lists also force you into higher daily volume, which means more mailboxes, more warm-up, and more surface area for something to go wrong. Sending 400 emails a day across eight mailboxes is a much bigger operational commitment than 80 a day across two.

Benchmarks by industry and buyer

These ranges are directional, drawn from aggregate cold outbound patterns rather than a controlled study. Use them to work out whether you're in the right neighbourhood, not to grade yourself to the decimal point.

Two patterns are worth noticing here. First, company size drives more variance than vertical does — a 15-person marketing agency and a 15,000-person bank behave completely differently even when you're selling the same thing. Segment your benchmarks by headcount band before you segment by industry. Second, a low reply rate is not automatically bad economics. A 2% reply rate into enterprise IT with $80,000 annual contracts beats a 12% reply rate into SMB at $200 a month, every time. Judge campaigns on pipeline per thousand leads, not on reply rate alone.

Diagnosing a bad number

When a campaign underperforms, work through the possible causes in roughly this order — it's the order of how much impact each one typically has.

• Reply rate near zero, bounce rate normal? Suspect inbox placement before copy. Nobody replies to mail they never saw. Run a seed-mailbox placement test and check your spam score before you rewrite a single subject line — the 15 common causes of cold email going to spam covers the usual suspects.
• Bounce rate above 5%? Your list is stale or scraped. Fix that first, because bounces are simultaneously dragging down deliverability for every other campaign running from the same domain.
• Replies coming in but 80%+ are "not a fit" or "wrong person"? That's a targeting problem, not a copy problem. Your filters are letting in companies that structurally cannot buy from you. Tighten the ICP.
• Emails clearly being read but almost no replies, positive or negative? You're being read and ignored, which is a relevance-and-offer problem. Usually it means you're describing your product instead of a problem the reader has already named out loud.
• First email gets replies, steps two through four get nothing? Your follow-ups say "just bumping this to the top of your inbox." Each step needs a new reason to respond: a different angle, a specific proof point, a smaller ask, or a graceful exit.

How much data before you trust the number

Reply rates are low-probability events, which makes them extremely noisy at small volumes. At a 5% baseline, a 200-lead test produces about 10 replies. Get 14 instead and it looks like a 40% improvement; in reality it's well within normal variance and you've just taught yourself something false.

• Don't read anything into fewer than about 30 total replies.
• For A/B tests, plan on 400–500 leads per variant at a 5% baseline before you call a winner.
• Test one variable at a time — subject line, or opening line, or CTA. Not all three.
• Test big swings, not word choices. Changing the offer or the segment moves reply rate by whole percentage points. Changing "Quick question" to "Quick one" moves nothing you can measure.

Also watch the trend across weeks, not just the campaign total. A reply rate that starts at 7% and decays to 2% over three weeks on a list of consistent quality is a deliverability signal — your mailboxes are drifting toward the spam folder — not a creative signal. Rewriting copy at that point makes things worse, because you'll conclude the new copy failed too.

Measuring this in Sendvanta

Sendvanta reports reply rate per delivered lead by default, separates genuine human replies from auto-responders and security scanners, and treats verified site visits as the engagement signal rather than raw opens and scanner-triggered clicks. Layered health scoring across mailbox, domain, DKIM, and IP will slow or pause sending when risk thresholds are crossed — which matters for this topic specifically, because the most common cause of a collapsing reply rate isn't your writing, it's placement quietly degrading while the dashboard still cheerfully reports emails as "sent." You can connect your own Gmail, Microsoft 365, or SMTP mailboxes on the free plan (1,000 active leads, 3,000 emails a month, no credit card) and see what your real numbers look like once the machines are filtered out. More detail on the reporting and deliverability side is on the features page.

Benchmarks compiled from aggregate B2B cold outbound patterns and reviewed in February 2026; treat them as ranges to orient against rather than targets to hit exactly.