Cold email benchmarks for B2B: Reply rates, positive replies and meetings per 1,000 prospects

A typical business-to-business cold email reply rate runs from 0.45% to 3.43% of total emails sent. Those are the averages Belkins reported from 7.5 million emails in 2025 and Instantly reported in 2026, with 5.5% as Instantly's top quartile. Further down the funnel, one agency's 1,533,399 sends produced a 1.9% positive reply rate and a 0.8% meeting-book rate. Check the denominator before you compare any figure.

What is a good cold email reply rate?

On total sends, Belkins measured a 0.45% average reply rate across 7.5 million emails in 2025, while Instantly reported 3.43% average and 5.5% top quartile. Both count all replies over emails sent, so treat 0.45% to 3.43% as the average band for that method and 5.5% as an upper-quartile mark.

In Belkins' 2025 data, reply rates on total sends fell from 0.72% at companies with 0 to 10 employees to 0.22% at companies with 10,000+ employees.

Cold email reply rate tiers, Instantly 2026 report

Item	Value
Elite (exceeds)	10%
Top quartile	5.5%
Average	3.4%

Source: Instantly, Cold Email Benchmark Report, 2026

Lavender's sample of 231,818 cold emails, pulled as of February 4th 2026, uses a different numerator. It counts positive and negative responses and excludes out-of-office and bounce notifications. It is filtered by recipient department, seniority and industry, so compare a segment there with your own segment, never with a blended average.

We think denominators matter more than vendors. Belkins notes that a 5% rate against openers and 0.45% against sends can describe the same campaign. For example, 1,000 sends at 3.43% is 34.3 replies, and 1,000 sends at 0.45% is 4.5 replies.

Reply rate by segment

Belkins' 2025 data, all on total sends:

• Seniority: founders and owners replied at 0.57%, C-level executives at 0.42%, and VPs at 0.32%.
• Company size: 0 to 10 employees replied at 0.72%, 11 to 50 at 0.49%, and 10,000+ at 0.22%.
• Industry: Food & Beverage, Education, and Government were the standout industries in 2025.
• Timing: among working days, Thursday and Wednesday reached 0.48% each, and morning sends from 8 AM to 12 PM reached 0.54%. Instantly found Wednesday highest.

The definition template

This is the template we use to report a reply rate. Fill it in before you compare your number with any row above.

REPLY RATE DEFINITION
Numerator:     [all replies | human replies only | positive only]
Auto-replies:  [included | excluded]
Bounces:       [included | excluded]
Denominator:   [total emails sent | unique prospects | unique openers]
Steps counted: [step one only | all sequence steps]
Segment:       [seniority | company size | industry | none]
Period:        [start date to end date]
Sample size:   [number of emails or prospects]
Comparable to: [benchmark row with the same numerator and denominator]

If the numerator or the denominator differs from the benchmark's, the comparison does not hold. Pick another row or restate your own figure.

Why cold email benchmarks disagree by a factor of ten

Published reply rates do not measure one thing. One roundup lined up three reports that put the average cold email reply rate at 3.43%, 8.5% and 2.09%. In our reading, almost all of that spread comes from three choices made before anyone counts a reply: the denominator, the reply definition and the sample.

The denominator

The previous section showed how one campaign can carry two very different rates depending on whether openers or total sends sit underneath. Delivered emails are a third base. Most reports don't separate delivered emails from sent emails, so a deliverability problem quietly lowers a sends-based rate. Instantly's method is to exclude hard bounces from the denominator for accuracy.

The reply definition

At one company, the team's reported figure included out-of-office messages and unsubscribes. Leadership reported a 12% reply rate to the VP. The actual positive reply rate was closer to 2.3%. Same mailbox, two stories. Instantly's classification splits replies into Positive, referral, objection, not now, unsubscribe, out of office, which gives you the categories to report separately.

Time window matters too. One sender's data shows 40% of replies come after day 10, so a seven-day cutoff undercounts.

The sample

The core question is whether a report covers agency sends to purchased lists, or in-house SDRs emailing qualified accounts. SDR stands for sales development representative. An in-house team working a named account list and an agency working net-new contacts are not running the same test.

A worked example

These are round example inputs, not a benchmark. For example, 1,000 sends at a 0.04 reply rate gives 40 replies. Suppose 0.3 of those 40 replies are positive, which gives 12 positive replies. If you report "4%" alone, the reader cannot tell which of these numbers you meant.

A definitions line for every report

Put this at the top of any reply rate you share, internally or externally, so the figure can be compared:

Denominator: [sent | delivered (hard bounces excluded) | opened]
Replies counted: [all | positive | positive + referral]
Excluded: [out of office | unsubscribe | auto-reply | none]
Window: replies within [N] days of last step
Sample: [list source: CRM / purchased / scraped], [sender: agency / in-house]
Reply rate = replies counted / denominator x 100

Before comparing your number to any published benchmark, fill the same five lines for the benchmark. If you cannot, treat the comparison as directional only.

Are cold email open rates still meaningful?

Open rates still exist in most sending tools, but they no longer work as a denominator for a pipeline model. Treat them as a narrow testing signal, read inside one sender's own setup, and keep them out of anything that forecasts meetings.

Why one sender stopped tracking opens

Belkins said it moved away from opens because the rate tracking pixel was hurting deliverability throughout the industry in 2024. Once it removed the pixel, open rate data became unreliable once we stopped tracking it, so it rebuilt its reply rate on total sends. The pixel is a tiny image that loads when a message renders. Mailbox providers can read it as a bulk-sender marker. If removing it is part of your inbox placement work, the steps sit in our outbound deliverability guide.

Measurement is a second problem. Gong called open rates a controversial metric in cold email right now and noted that Those opens could be the prospect’s spam filters. Any open that a security scanner fires is noise in the denominator.

How an open-based denominator inflates a benchmark

This is an example, not a benchmark. For example, a campaign sends 1,000 emails and gets a 0.04 reply rate on sends, so 1,000 times 0.04 gives 40 replies. Suppose the open rate on those 1,000 sends is 0.3, so the open count is 300. In this example, dividing the 40 replies by the 300 opens makes the reported reply rate about 13.3%. The replies did not change. Only the denominator shrank, and in our reading that is the mechanism behind the higher numbers in older open-based benchmarks.

The narrower use that remains

Opens can still answer one question: which subject line got more attention, compared inside the same sender, list and tracking setup. Executives give that question little time. Gong reported that from the inbox view they typically take less than three seconds to decide whether to open an email. It also found that email open rates decline as subject lines get longer, with 1, 4 words performing best.

We read that as a hypothesis to test, not a rule to adopt. A short test plan keeps opens in their place:

SUBJECT LINE TEST (opens as a relative signal only)
Variant A: subject of 1 to 4 words
Variant B: current subject line
Same sender domain, same list segment, same send window
Tracking: identical on both variants, or off for both
Primary metric: replies / delivered
Secondary metric: opens / delivered (direction only)
Decision rule: pick the variant with more replies;
  use opens only to break a tie
Do not compare open rates across senders or tools

Report opens as direction, never as a rate you hand to finance. The pipeline number runs on replies and meetings per delivered email.

Positive replies and meetings are the numbers to benchmark

A reply count mixes interest, refusals and requests to stop. A meeting count does not. That is why the useful benchmark sits further down the funnel than most reports go.

Across one agency's quarter of sends, the rate falls from 4.7% for all replies to 1.9% for positive replies and 0.8% for meetings booked.

Reply, positive reply and meeting rates on 1,533,399 cold emails

Item	Value
All replies	4.7 % of emails sent
Positive replies	1.9 % of emails sent
Meetings booked	0.8 % of emails sent

Source: oneaway.io, cold email benchmark, 2026

One agency's quarter shows the spread. Across 1,533,399 cold emails sent for 17 clients, the blended reply rate was 4.7%, the positive reply rate was 1.9% and the meeting-book rate was 0.8%.

Each stage loses more than half of the one before it. In our reading, a team that reports only the first row can look healthy while the third row stays flat.

Emails per meeting

Gong analyzed 28M+ cold emails and found that the average rep has to send 344 cold emails to land one meeting. Expressing the benchmark as emails per meeting puts it in the same unit as capacity planning: sends per rep per week, divided by that figure, gives meetings per rep per week.

The average hides a wide spread. In the same data, the top 10% of reps land 8.1x more meetings than average reps, and the top 25% land 4.3x more. We think a team should benchmark against its own reps first, because that gap is larger than most differences between published averages.

Segment by seniority before comparing

Audience changes the baseline. Gong Labs data show that C-level executives are 30.2% less likely to reply to cold emails than non-executives. A sequence aimed at chief officers should carry its own target, not the blended one.

Worked example

These are round example inputs, not a benchmark. For example, with 1,000 prospects and a 0.04 reply rate, a campaign gets 40 replies.

Use one funnel record per campaign, with every rate on the same denominator:

Campaign:            ______
Segment (seniority): ______
Prospects contacted (delivered): P
All replies:         R      Reply rate          = R / P
Positive replies:    Q      Positive reply rate = Q / P
Meetings booked:     M      Meeting rate        = M / P
Emails per meeting:          = P / M
Positive-to-meeting:         = M / Q

Meetings that pass qualification feed your sales-qualified lead (SQL) count, and from there the win rate benchmarks take over.

How many emails should a cold email sequence have?

Most of the reply volume comes early. Instantly found that 58% of all replies are generated from step one, and that remaining follow-ups contribute 42% of total replies. The first email sets your ceiling. The follow-ups still carry a large share of the result.

The published guidance on length points in two directions.

The two can be reconciled. The blog also notes that the first follow-up adds 40-50% more replies. In our reading, steps two and three earn their place on almost any list.

The sequence ends before the deal does

Sopro pointed out a mismatch: a share of deals takes over a year to close - yet most outreach sequences last just 4-6 weeks.

We think the practical answer is not a longer cold sequence. It is a short, dense cold sequence followed by a slower track for prospects who did not reply but still fit your ideal customer profile. Check your own cycle against sales cycle length benchmarks before you decide how long that second track runs.

A sequence plan you can copy

Fill in the day numbers from your own testing. Track replies by step so you can see where your own split of first-touch and follow-up replies lands.

COLD SEQUENCE (active phase)
Step 1  Day [x]   Problem-first opener, one call to action
Step 2  Day [x]   Short reply-style bump on the same thread
Step 3  Day [x]   New angle: case study or proof point
Step 4  Day [x]   Different pain point, same persona
Step 5+ Day [x]   Only if it adds new value; otherwise stop

EXIT RULES
- Positive reply      -> book meeting, remove from sequence
- "Not now"           -> move to nurture, set a revisit date
- No reply, good fit  -> move to nurture track
- Unsubscribe/bounce  -> suppress

NURTURE TRACK (long phase)
- One touch every [x] weeks, new content each time
- Re-enter cold sequence on a trigger: role change,
  funding, hiring, or intent signal
- Review the track against your sales cycle each quarter

REPORTING PER STEP
replies_step_n / total_replies = share of replies by step
positive_replies_step_n / delivered_step_n = positive rate by step

Cut any step whose positive rate stays flat after a fair sample. Length should follow the per-step numbers, not a template count.

Does personalization raise cold email reply rates?

Personalization raises reply rates in the published data, but the size of the lift depends on who receives the email and what the personalization points at. It analyzed more than 30,000 prospecting emails from more than 250 companies and sorted them into four personalization types.

Each figure in that table is a lift against Gong's own baseline for that segment, not an absolute reply rate. Gong found little correlation between higher reply rates and company mentions with lower-level buyers, so the company-based result does not carry down the org chart. The activity-based approach is the only one Gong tied to meetings as well as replies, and those leads are often already warmer, which in our reading inflates its lift.

Relevance limits all four. In the same study, 87% of buyers say the sales emails they receive don't address a relevant challenge facing their organization.

Personalization does not offset a pitch

The same analysis put the ideal length at 100 words or fewer. We think a personalized opening followed by a product pitch gives back most of what the opening earned.

Reading score-based lift figures

Lavender publishes lift by segment for emails that score well against writing best practices. Its definition matters: lift is the % change going from Average Reply Rate to 90+ Score Reply Rate. That is a relative change within one segment, so a large lift on a low-replying segment can still produce fewer replies than a small lift on a high-replying one. Compare the absolute 90+ reply rates, not the lift column, when you pick segments.

To test personalization on your own list, split by seniority before you split by message:

Test: personalization type by seniority
Segment A: individual contributors and managers
  Variant 1: individual-based opener
  Variant 2: no personalized opener
Segment B: directors and above
  Variant 1: company-based opener
  Variant 2: no personalized opener
Hold constant: body copy, length, call to action, send days
Measure per segment: reply rate on delivered, positive reply rate, meetings booked
Decision rule: keep the variant only if positive replies rise, not just total replies

How deliverability rules change cold email results

Every benchmark in this piece assumes the mail reached an inbox. When it does not, reply rates fall for reasons that have nothing to do with copy, so the sender conditions come first.

The conditions a benchmark assumes

Instantly's guide adds that bulk senders must meet 2024 Gmail and Yahoo standards. Bounces come in two kinds: a soft bounce is caused by transient errors, such as a full mailbox, while hard bounces are permanent errors, such as addresses that do not exist.

Why the reply rate drops when mail misses the inbox

The usual formula divides replies by delivered emails, and the guidance is to exclude hard bounces from the denominator. In our reading, that still leaves a gap: mail filtered to spam counts as delivered, so it stays in the denominator while producing almost no replies.

reply rate = replies / delivered emails x 100
delivered emails = sends - hard bounces
inbox reached = delivered emails - messages filtered to spam

For example, and not as a benchmark, take 1,000 sends where a 0.3 share lands in spam and inbox recipients reply at 0.04. In this example, spam takes 1,000 × 0.3 = 300 messages, which leaves 1,000 − 300 = 700 in the inbox. In this example, the inbox messages produce 700 × 0.04 = 28 replies. Divided across all sends in this example, 28 replies over 1,000 sends is 0.028, or 2.8%.

The copy did not change. The list did not change. In this example, the reported rate fell from 4% to 2.8% because three in ten messages never had a reader. One check from the same guide helps here: drops isolated to Microsoft or Google can signal throttling, so split open and reply rates by mailbox provider before rewriting anything.

What outreach tools automate

Outreach tools now handle much of this setup. A 2022 review by InStream Group described its automatic e-mails and follow-ups and sending one at a time in the hourly intervals specified by us.

For the full technical setup, see our outbound email deliverability guide.

What this page does not cover

This page does not cover the full technical setup of sending domains and authentication, which sits in our outbound email deliverability guide, or what happens after a meeting is booked, such as win rates and sales cycle length. It does not compare or recommend any outreach tool or agency, and it is not legal advice on email sending rules.