What Is a Good Cold Email Reply Rate? 2026 Benchmarks and How to Read Them
Quick answer
It depends on what you divide by. Published 2026 figures run from 0.74% (Smartlead's median sender, per person contacted) to 4.5% (Hunter's average sequence, per recipient), and Instantly reports a 3.43% average per email sent. Compare your rate only with a benchmark that uses the same formula, take auto-replies out, and judge the campaign on interested replies and meetings rather than on replies alone. Facts checked 14 November 2026.
Key takeaways
- Read the denominator first. Smartlead's 0.74% median divides by people contacted, while Instantly's 3.43% average divides by emails sent.
- The same campaign can look weak or excellent depending on the formula. In our worked example, one set of replies reads as 1.0%, 2.1% or 3.0%.
- Raw tool counts include out-of-office notices and bounce messages. Take them out before you compare.
- Judge a campaign on interested replies and booked meetings, not on the reply rate alone.
- When the numbers are low, check the offer first, then the positioning, the copy, the data and the sending setup.
Benchmarks- 1
What was it divided by: emails sent, people contacted or sequence recipients?
- 2
Are out-of-office notices and bounce messages included?
- 3
Is it a median or an average?
- 4
Whose campaigns does the benchmark measure?
- 5
How many of the replies turned into interested leads and meetings?
Your first month of cold email is done. The sending tool says 2.8%, a blog post says anything under 5% is weak, and your business partner wants to know whether the campaign is failing. Nobody in the room knows what the 2.8% was divided by.
This guide sets out the four 2026 benchmarks we checked, why they disagree, a worked example that shows how much the formula matters, and what to look at when the numbers are low.
What is a good cold email reply rate in 2026?
It depends on what you divide by. Published 2026 figures run from 0.74% per person contacted to 4.5% per sequence recipient, so the first column to read in any benchmark is the denominator.
| Source | Reply rate | Divides by | Data | Period |
|---|---|---|---|---|
| Smartlead, State of Cold Email 2026 | Median sender 0.74% (one reply per 135 contacts). Top 25% 1.43%, top 10% 2.63%, top 5% 3.85% | People contacted | 850 million+ emails, senders with 500+ sends | Jan to Jun 2026 |
| Instantly, Cold Email Benchmark Report 2026 | Average 3.43%. Top 25% 5.5%+, top 10% 10.7%+ | Emails sent, with all replies including follow-ups | Instantly’s platform data | 1 Jan to 18 Dec 2025 |
| Hunter, State of Email Outreach 2026 | Average sequence 4.5%, sales outreach 3% | Recipients of a sequence (45 replies per 1,000) | 31 million emails | 2025 |
| SmartReach, State of Cold Email 2026 | Agency cohort about 10% reply, 2.2% positive, 1.2% meetings booked | Not defined for positive replies or meetings | 40 million emails from its own customers | Jan to Jun 2026 |
Match your own formula to the source. Rates per email sent belong next to Instantly’s report, and rates per person contacted belong next to Smartlead’s bands.
Why do the published benchmarks disagree?
They divide by different numbers, count different kinds of reply, use medians or averages, and measure different senders. The denominator alone can triple a rate, as the worked example below shows.
- The denominator. Instantly divides by emails sent, Smartlead by people contacted, and Hunter by recipients of a sequence. Follow-ups add to emails sent but not to people, so the same campaign looks better per person than per email.
- Median or average. Smartlead reports the sender ranked halfway down its list. Instantly and Hunter report averages, and a handful of very strong senders can pull an average well above the middle.
- What counts as a reply. Out-of-office notices and “she has left the company” messages are replies from a machine. Ask whether any figure includes them.
- Whose campaigns. Each source measures its own users. SmartReach’s figures come from its own customers. Hunter’s 4.5% is its average across sequences, while its sales outreach average is 3%.
- How long the sequence is. Instantly found that 58% of replies come from the first email. Every extra follow-up adds sends faster than it adds replies, which lowers a per-email rate and raises a per-person one.
How much does the denominator change the number?
Enough to turn an average campaign into a top one, without a single reply changing. Here is a made-up example.
Lena runs new business at a 60-person design agency. She emails 800 people, three emails each, so 2,400 emails go out. She gets 24 replies, and 7 of them are out-of-office notices.
- Per email sent, 24 divided by 2,400 is 1.0%.
- Per person contacted, 24 divided by 800 is 3.0%.
- Per person, counting only real people, 17 divided by 800 is about 2.1%.
Against Instantly’s 3.43% per-email average, Lena’s 1.0% looks weak. Against Smartlead’s per-person bands, her 3.0% sits between the top 10% (2.63%) and the top 5% (3.85%). Counting only people, her 2.1% still sits between the top 25% and the top 10%. Same campaign, same replies, three verdicts.
Write the formula next to any rate you quote.
- Reply rate per email sent is replies divided by emails sent.
- Reply rate per person contacted is replies divided by unique people who got a first email.
- Human reply rate is the same, with auto-replies and bounce notices taken out of the reply count.
Why do raw reply counts overstate?
Because sending tools count every message that comes back, and many of them come from machines: out-of-office notices, “no longer with the company” messages and mailbox-full warnings. On a large campaign these can make up a real share of the total.
Two checks keep the number honest.
- Read a sample by hand. Take the last 100 replies, mark which came from a person, and apply that share to the raw rate. It is a rough correction, but it tells you how far the tool’s figure sits from reality.
- Read the positive tags too. Tools tag replies as interested from the words in them, and a question about price or a request to “send more info” often gets the same tag as a firm yes. A tag is a signal, not a meeting.
If you are judging an agency, ask how it counts replies, what it calls positive, and whether a person checked.
What should you judge a campaign on instead?
On what the replies turn into. A campaign with a modest reply rate that produces interested buyers beats one with a high reply rate made of unsubscribes and “not now”.
Track three numbers over the same dates:
- Interested replies, read by a person, per hundred or thousand people contacted.
- Booked meetings, counted when the time is on the calendar.
- Held meetings, counted when the call actually took place.
For what belongs in that middle number, and how to write it into a contract, see what counts as a qualified meeting.
Which reply rate should worry you?
Zero is the clearest signal, and it usually points at delivery, not copy. No copy is bad enough to stop every out-of-office notice, so a campaign that has reached a few thousand people without a single reply of any kind is almost always a campaign whose emails are not arriving. Check the sending domains, the inbox health and the bounce rate before you touch a word of the email.
For an outside reference point, Smartlead’s median sender gets one reply per 135 contacts.
What actually raises a reply rate?
Work from the top of this list down. In our view, the order of what to check is offer, positioning, copy, data and sending setup. The last three are the easiest to change, which is why teams tend to start there.
- Offer. For a services firm, “we build software” is not an offer. One specific project you have shipped before, aimed at a company that needs it, is.
- Positioning. Describe the problem in your buyer’s words, not in your capabilities deck.
- Copy. Since 58% of replies come from the first email, according to Instantly’s report, that email deserves the most work. One clear idea for one company reads better than a list of services.
- Data. Who is on the list matters more than how many. A smaller list of companies that clearly need what you sell beats a large list bought by industry code.
- Sending setup. Google asks every sender to authenticate with SPF or DKIM and keep spam complaints under 0.3% (Google sender guidelines). It is the easiest layer to fix and the least likely to be the real problem once the basics are in place.
If you are hiring help, the reply and meeting numbers to ask for are part of the checks in our guide to choosing a lead generation agency. If you are weighing a hire instead, the in-house SDR or agency comparison puts both on a cost per meeting.
The bottom line
A reply rate is only useful when its formula travels with it. State the denominator, state whether auto-replies are in, line yourself up against sources that use the same formula, and judge the campaign on interested replies and meetings rather than replies.
If you want a second read of your own numbers, book a call and bring the raw counts.
Vergelead at a glance
Where Vergelead fits
Vergelead runs idea-led outbound for software development companies and creative agencies: it researches each prospect, finds a gap your team could close, and pitches that one idea by cold email and LinkedIn, so qualified meetings land on your calendar.
- What it is
- A cold email and LinkedIn outbound agency in Sofia, Bulgaria, founded by Borislav Nekezov.
- Who it is for
- Software development and IT services firms of about 30 to 300 staff, and creative and digital agencies, selling into the UK, the EU, the Nordics and the US.
- How it works
- Every first email pitches one build or project idea drawn from that prospect’s own business. A person approves every idea before it is sent, and a person on our team answers replies and books the meeting.
- What you own
- The sending domains, inboxes, sequences and lead lists built for you. On request, we transfer all of them, lead lists as a CSV with reply status, within 14 days of the final invoice.
- Sending setup
- Google Workspace inboxes in the client’s name. Every lead, reply and meeting shows in a client portal that syncs live.
Facts checked 14 November 2026
Frequently asked questions
Is a 1% cold email reply rate good?
It depends on the denominator. Per person contacted, 1% is above the median sender in Smartlead's 2026 data, while per email sent it is below Instantly's 2026 average. The benchmark table in this guide has both, with sources. Either way, look at how many replies become interested leads before you change anything.
Should I count replies per email sent or per person contacted?
Both are fine as long as the formula is stated. Per person always comes out higher once follow-ups go out, because follow-ups add emails but not people. Per person is steadier, since it doesn't change when you add a step to the sequence.
Do auto-replies count in the reply rate?
Many sending tools count them, so ask. Out-of-office notices and 'no longer with the company' messages are replies from a machine, not interest. Say whether auto-replies are in any rate you quote, and take them out before comparing with a benchmark.
How many cold emails per booked meeting should I expect?
It depends on the list, the offer and the market far more than on the copy, so any single figure will mislead you. Track it on your own campaigns: emails or people contacted, then interested replies, then booked meetings, each with the same date range.
Which reply rate should worry me?
A campaign that gets zero replies of any kind, not even out-of-office notices, usually has a delivery problem rather than a copy problem. Check the sending domains, inbox health and bounce rate first.
How many replies should be positive?
Define positive first. Some tools tag any non-negative reply as interested, including questions about price. A tag is a signal, not a meeting, so have a person read the replies before you count them as leads.