What is a good cold email reply rate?
It depends on what is being counted. Benchmarks for cold emails sent in 2025 run from 0.45% at Belkins to 4.5% at Hunter, and both can be right because they count replies differently and divide by different bases. A good rate is one that beats a benchmark measured the same way you measure yourself.
If your tool counts replies per person contacted, Smartlead's percentiles are the closest fit: its median sender gets replies from 0.74% of contacts, the top 25% reach 1.43% and the top 10% reach 2.63%. If your tool counts replies per email sent, Instantly's 3.43% average and Belkins' 0.45% are the two figures here on that base, and the next sections explain why they sit so far apart.
How do the published benchmarks compare?
Their headline rates range from 0.45% to 4.5%, and each comes with its own definition, sample and period. Every one is the vendor's own data, describing its own users or clients, checked on 20 October 2026.
| Source | Reported rate | What counts as a reply, divided by what | Sample | Data period (page date) |
|---|---|---|---|---|
| Belkins (outbound agency) | 0.45% average | Unique replies, excluding auto-replies and bounce notifications, divided by emails sent | 7,530,489 emails and 34,393 replies from its client campaigns | January to December 2025 (updated 26 June 2026) |
| Smartlead (software vendor) | 0.74% median sender; top 10% 2.63% | Replies divided by contacts; auto-replies not mentioned; only senders with at least 500 emails | More than 850 million emails | Not stated (no date shown) |
| Instantly (software vendor) | 3.43% average; top 25% 5.5%; top 10% 10.7% or more | All replies, including replies to follow-ups, divided by emails sent; auto-replies not mentioned | Billions of interactions across thousands of workspaces, by Instantly's description | 1 January to 18 December 2025 (updated 12 January 2026) |
| Hunter (software vendor) | 4.5% average; 3% for sales sequences | Replies per recipient of a sequence; auto-replies not mentioned | 31 million emails sent by Hunter users | 2025 (2026 report, no date shown) |
| Saleshandy (software vendor) | 4.5% average | Total replies received divided by emails delivered; replies are sorted into interested, not interested, auto-reply, out of office and do not contact | 30,314,409 emails and 40,006 sequences | The method section says July to October 2026; the reply-rate section says January to June 2026 (updated 7 October 2026) |
Don't average these numbers. They use different bases and different senders, so a mean of them describes no real campaign. Saleshandy's 4.5% needs extra care: on the same page, its follow-up emails reply at 0.37% to 0.80% each, far below that average.
Why do cold email benchmarks disagree?
Because each source makes four choices differently. Smartlead's own explanation is that much of the gap comes from what gets counted and whether the sample is one agency's campaigns or a whole platform.
- The base. Per email sent (Instantly, Belkins), per email delivered (Saleshandy) or per person (Hunter, Smartlead). A three-email sequence to 100 people is 300 emails. If six people reply, that is 6% per person and 2% per email, from the same campaign.
- What counts as a reply. Belkins removes auto-replies and bounce notifications. Instantly, Hunter and Smartlead don't say whether out-of-office messages count, and Instantly notes that Friday brings its highest volume of auto-replies.
- Average or median. Smartlead reports the median sender because, it says, a few very high performers can pull an average up. Belkins, Instantly, Hunter and Saleshandy report averages.
- Who is sending, and why. Belkins reports its own agency campaigns, mostly to US contacts (over 5.6 million of its emails). Software platforms pool every user. Hunter's 4.5% includes digital PR sequences, which average 13%, and headhunting at 7.5%, while its sales sequences average 3%.
Here is one made-up campaign scored five ways. You email 200 people three times each: 600 emails sent, 10 bounced, 590 delivered. You get 12 replies, of which 3 are out-of-office messages and 3 are positive.
| Method | Sum | Reply rate |
|---|---|---|
| All replies per person | 12 / 200 | 6.0% |
| Human replies per person | 9 / 200 | 4.5% |
| All replies per email delivered (Saleshandy's wording) | 12 / 590 | 2.0% |
| All replies per email sent (Instantly's wording) | 12 / 600 | 2.0% |
| Human replies per email sent (Belkins' method) | 9 / 600 | 1.5% |
Hunter and Smartlead divide by people but don't say how they treat auto-replies, so their method could match either of the first two rows. The same 12 emails in your inbox produce rates from 1.5% to 6.0%. Only 3 are positive: 1.5 positive replies per 100 people.
Method changes matter too. Belkins' earlier studies divided replies by the people who opened an email; its 2026 study divides by emails sent. In Belkins' words, "A 5% reply rate against openers and a 0.45% reply rate against total sends can describe the same campaign."
Which benchmark should you compare yourself with?
The one that counts the way your tool counts. Check your tool's definition first, then use this guide:
| If your tool reports | Compare with | Benchmark |
|---|---|---|
| Replies per person contacted | Smartlead percentiles | Median 0.74%, top 25% 1.43%, top 10% 2.63%, top 5% 3.85% |
| Replies per person, sales sequences only | Hunter, sales use case | 3% average |
| Replies per email sent, auto-replies removed | Belkins | 0.45% average |
| Replies per email sent, all replies | Instantly | 3.43% average; 5.5% for the top 25% |
Then adjust for your market. Belkins' 2025 data runs from 0.72% for companies with up to 10 employees to 0.22% for those with 10,000 or more, and from 0.57% for founders and owners to 0.32% for VPs. Where a source publishes a segment close to yours, compare against that segment.
Why track positive replies instead of all replies?
Because a reply can be a no. Hunter's 2026 report says that among senders who get a reply, most receive negative replies. In Belkins' send-time study, Wednesday tied for the highest weekday reply rate in 2025 (0.48%) but had the lowest meeting rate of any weekday (0.1%), and Belkins suggested some send windows bring more brush-offs and requests to be removed, though its meeting rates are hard to square with the 1,200-plus appointments it reports from 7.5 million emails.
Positive reply rate is harder to find in public benchmarks. Saleshandy is the only one of the five that publishes it: an average of 3% to 5% of delivered emails. That sits close to its 4.5% total reply rate on the same page, so read both figures with care. Hunter's guide to reply rate calls positive reply rate a crucial metric, and counts a reply as positive when the recipient is open to moving forward.
Some practitioners already work this way. Jed Mahrle, founder of the outbound agency Practical Prospecting, splits a campaign into reply rate (fixed with copy), positive reply rate (fixed with the offer) and meeting conversion (fixed with follow-up). Mahrle's April 2026 worked example: 500 people contacted, a 2% reply rate gives 10 replies, 20% of those positive gives 2, and half of those booking gives 1 meeting. Mahrle adds that most people are in the range of 1,000 to 5,000 leads per meeting booked. For the step after a positive reply, see our guide to booking more sales meetings.
How to track positive replies and meetings per 100 contacted
This is our recommendation: report three numbers for every campaign, each per 100 people contacted. Counting people instead of emails means a longer sequence doesn't change the base.
Count people who received at least one email
Use unique people as the base, after removing bounces; Hunter's guide also subtracts bounces, though its formula divides by emails, not people. If you use Siela's AI SDR, each run's leads are saved to their own CRM group with the reason each was picked, which gives you a starting list; count only the people who were actually emailed.
Sort every reply into a bucket
Positive (interested or asking a question), not now, no, auto-reply or out of office, and referral. Mahrle's team sorts replies into four similar buckets and follows up referrals and replies saying the person has left. Leave auto-replies out of your reply count.
Report replies, positive replies and meetings per 100
Mahrle's example works out at 2 replies, 0.4 positive replies and 0.2 meetings per 100 people contacted (the example's numbers divided by five). Treat that as a sense of scale from one agency. Log each meeting in your CRM against the campaign that produced it.
Wait for enough volume
Small samples swing. Mahrle tests messaging at around 500 emails per variation, and says there is no strict rule.
Fix the weakest number first
Few replies of any kind point to deliverability, the list or the copy. Replies without positive ones point to the offer. Positive replies without meetings point to how fast and how often you follow up. Our guide to cold email sequences covers the copy.
Frequently asked questions
Do out-of-office replies count in cold email reply rate?
It depends on the source. Belkins excludes auto-replies and bounce notifications from its 0.45%. Saleshandy divides total replies by delivered emails and sorts replies into categories that include auto-reply and out of office. Instantly, Hunter and Smartlead don't say either way in their methods. For your own tracking, leave auto-replies out so the number counts people.
Is cold email reply rate measured per email or per person?
Both are in use. Instantly and Belkins divide replies by emails sent, Saleshandy by emails delivered, and Hunter and Smartlead by people. The same campaign looks lower per email, because every follow-up adds to the base: 6 replies from 100 people sent 3 emails each is 6% per person and 2% per email.
What is a positive reply rate?
It is the share of emails or people that get an interested reply, such as a request for more information or a meeting. Saleshandy defines it as interested replies divided by emails delivered. Hunter's guide counts a reply as positive when the recipient is open to moving forward. Few public benchmarks report it, so track your own.
Why is Belkins' cold email reply rate so much lower than other benchmarks?
Belkins divides unique replies by every email sent and removes auto-replies and bounce notifications, the strictest of the five methods compared here. It also reports only its own agency's client campaigns, mostly to US contacts. Its earlier studies divided by opened emails, and Belkins says that change of method is why its 2025 figures read lower than past editions.
How many people do you need to contact to book one meeting?
There is no reliable public benchmark. Jed Mahrle's April 2026 example, built on a 2% reply rate, 20% positive replies and half of those booking, gives one meeting per 500 people, and Mahrle says most people are in the range of 1,000 to 5,000 leads per meeting. Measure your own meetings per 100 people contacted.
Sources
- Belkins: What are B2B cold email response rates? (2026 study) (updated 26 Jun 2026)
- Belkins: What is the best time to send B2B emails? (2026 study) (updated 26 Jun 2026)
- Smartlead: What is a good cold email response rate? (checked 20 Oct 2026)
- Instantly: Cold Email Benchmark Report 2026 (updated 12 Jan 2026)
- Hunter: The State of Email Outreach 2026 (checked 20 Oct 2026)
- Hunter: Improve Your Reply Rate (cold email guide) (published 4 Dec 2023, checked 20 Oct 2026)
- Saleshandy: I Analyzed 30.31 Million Cold Emails (Dhruv Patel) (updated 7 Oct 2026)
- Jed Mahrle, Practical Prospecting #145: Why You Should Send More Emails (published 19 Apr 2026)
- Jed Mahrle, Practical Prospecting #129: The 3 Phases of Every Cold Email Campaign (published 19 Oct 2025)
- Jed Mahrle, Practical Prospecting #152: How to Convert More Meetings From Email (published 21 Sep 2026)
- Jed Mahrle, Practical Prospecting: Update (newsletter has a new home) (published 19 Jun 2025)