Cold Email Reply Rate Benchmarks and What Counts as Good

A realistic cold email reply rate in 2026 runs from under 0.5% to about 5%, and your position in that band depends first on how the number is counted.

Two large 2026 reports put the typical sender under 1%. One agency's study of 7.5 million emails from its own 2025 campaigns found that 0.45% of emails sent got a real reply, with auto-replies and bounces stripped out.

Smartlead's State of Cold Email 2026, built on more than 850 million emails, found a median of one reply per 135 people contacted, which is 0.74%.

Counted the way a sending platform shows it on a dashboard, every reply over every email sent, the numbers look better. Instantly's 2026 benchmark report puts the average at 3.43%, the top quarter at 5.5% or more and the top tenth at 10.7% or more.

So a dashboard reading of 3% to 5% is normal, and above 5% is good. Under 1% is ordinary for a cold list when replies are counted per person contacted, or per email with auto-replies stripped out.

Under 2% on a platform dashboard is, in Apollo's tiering, below average and usually points at deliverability, targeting or generic copy.

All four sources above are vendors. Three report their own platform or campaign data, and Apollo summarizes other studies, so read them all as vendor numbers. Below are six sources in one table, what each one actually measured, what moves the rate, and four of our own campaigns with raw counts.

Those campaigns come from our client work. Smirnov Consulting Group is a Prague-based B2B outbound lead generation agency that runs cold email and LinkedIn campaigns for founder-led B2B companies and books qualified sales calls.

Why do cold email reply rate benchmarks disagree so much?

Three choices change the result, and most pages quoting a benchmark mention none of them.

What counts as a reply. The agency study counts unique replies and drops auto-replies and bounce notices. Instantly counts all replies received, follow-up replies included.

Apollo's guide points out that a total reply count also carries out-of-office messages, unsubscribe requests and flat noes, none of which put anything in your pipeline.

What sits underneath. Instantly and the agency study divide by emails sent, so a four-step sequence to one person counts four times in the denominator. Smartlead divides by people contacted and reports the result as how many contacts it takes to earn one reply: 135 for the median sender.

Who is in the sample. Smartlead includes every sender on its platform with at least 500 emails in the first half of 2026, the careful ones and the careless ones. Instantly reports across thousands of workspaces for 2025. The agency study covers only that agency's own campaigns in 2025.

One detail is worth a second look. Dividing by people instead of emails should push a rate up, because each person receives several emails. Smartlead's per-person median still comes out far below Instantly's per-email average.

That leaves the sample, the definition of a reply and the choice of statistic to explain the gap: a median is the middle sender, while an average can be pulled up by a few senders with very high rates.

There is also a quieter trap: the same figure travels between pages. Woodpecker's 2026 statistics page shows 3.43% as well, but it quotes and links Instantly's report for it. Two pages, one dataset.

When you see a number repeated across several vendor blogs, check whether they measured it or borrowed it.

Six sources side by side

Every row comes from a company selling cold email or outreach software or services, or from a study run with one. The numbers can still be right, so weight them by how much they disclose. The agency study, Smartlead and Instantly state their sample and their denominator, and they deserve the most trust here.

What the table means if you run a B2B company: the only honest comparison is like for like.

If your tool divides all replies by emails sent, hold yourself to Instantly's column. If it divides replies by people contacted, use Smartlead's column. If you count unique replies per email sent, with auto-replies and bounces removed, the agency study is the closest match.

Mixing counting methods is how founders end up either panicking over a healthy campaign or celebrating a broken one.

The 8.5% figure is from 2019, and it was not sales email

The number most often quoted as the "average cold email response rate" comes from a Backlinko study run with Pitchbox. It analyzed 12 million outreach emails and found that 8.5% received a response. It was published in April 2019.

The campaigns in it were guest post pitches, link requests, expert roundups, sponsorships, reviews and similar content outreach. The study does not mention sales prospecting or B2B lead generation at all. A blogger answering a guest post pitch is a different event from a CFO answering a stranger who wants a meeting.

If an article presents 8.5% as a 2026 B2B cold email average, it is repeating a claim the study never made.

Which reply rate counts as good this year?

Here are the tiers each source publishes, in its own counting method. Read each column on its own. Smartlead states its tiers as contacts per reply (70, 38 and 26 for the top 25%, 10% and 5%), so its column shows those converted to percentages.

My practical reading of that table:

• On a platform dashboard that counts every reply per email: under 2% needs fixing, 3-5% is normal, 5% and up is good, 10% and up is rare.
• On a count of replies per person contacted (Smartlead's method): under 1% is typical for a cold list, about 1.4% or more puts you in the top quarter, and about 2.6% or more puts you in the top tenth.

A benchmark describes the average sender, and you are not the average sender, for better or worse. I made the same point in an earlier Journal post: one identical message pulled 7.5% replies when one B2B company sent it and 2% when another did. Only the sender changed.

The full story of that gap is about what prospects check before they answer. Here, the short version is enough: your profile, your proof and your list decide which side of the average you land on.

Who replies more: company size, seniority and industry

The agency study is the only source in the table that breaks its 2025 data down by who received the email. All of these rates sit on its strict 0.45% base, so compare them only with each other.

Two ratios matter more than any single cell. Founders and owners replied about 1.8 times as often as VPs. Companies with up to 10 people replied more than 3 times as often as companies with 10,000 or more.

For a founder-led company selling to other founder-led companies, that is good news: the person reading the email is often the one who can sign. If your list is mostly VPs at enterprises, your honest benchmark sits below the average line before you write a word.

Woodpecker's statistics page shows two more effects. Citing an analysis of 16.5 million emails by the same agency, it reports that campaigns under 50 recipients averaged 5.8% against 2.1% for large sends. The page gives the large-send cut-off as 500 recipients in one place and 1,000 in another, so treat the exact line loosely.

From Woodpecker's own data, cross-checked against other 2026 benchmarks, the page reports that advanced personalization reached roughly 17-18%, against roughly 7-9% with basic or no personalization. Both figures sit far above every average in the tables above, so read the gap between them, not the level.

I would not read the list-size gap as "send less". A 40-person list is usually hand-picked and hand-written, so much of that gap is probably personalization and fit. The useful lesson is that the same effort per contact does not survive a jump to thousands of contacts unless the list and the research scale with it.

Open rate, bounce rate and follow-ups: the numbers around the reply

Open rate. Woodpecker puts average opens at 27.7% to 44% and warns in the same breath that Apple Mail Privacy Protection preloads tracking pixels, which inflates the count.

Smartlead goes further and says open rate can no longer be trusted as a cold email metric. Treat it as background noise. Replies are the first number a human produced.

Bounce rate. The three sources that grade it agree. Instantly treats under 2% as the target, Smartlead calls 2% the danger line with a typical sender at 1.54%, and Woodpecker grades under 2% as good and under 1.5% as excellent.

My team checks every address before a campaign sends to it, and we keep bounces under 0.5%. Above 2%, your reply rate is being measured on mail that partly never arrived, so fix the list and the sending setup before you judge anything else.

Follow-ups. In Instantly's 2025 data, 58% of replies came from the first email and 42% from follow-ups, and it found 4 to 7 touchpoints to be the best-performing sequence length. So a sequence that stops after one or two emails gives up part of that 42%.

Across our campaigns, meetings tend to come from follow-up number three, provided it goes out after a proper pause. I cover the spacing in how I build a follow-up sequence.

Positive replies and booked calls: what four real campaigns show

Every number in the benchmark tables is an aggregate percentage. Here is what four of our campaigns looked like in raw counts, labeled by channel mix, because a reply rate means little until you know which channels produced it.

Campaign C ran cold email and LinkedIn side by side for one client in the same month, and I reported the two arms separately, so it gets two rows.

What happened after the calls: in A, 12 people showed and 3 signed. In B, 11 calls were completed and 2 contracts signed.

In C, 28 showed across both arms (16 from cold email, 12 from LinkedIn) and 4 contracts were signed, worth EUR 1,152,000 together. D is counted up to the booked call: 141 calls in 60 days.

Four things stand out.

No channel wins on reply rate every time. A and B, both LinkedIn-led, replied at 24% and 14%.

In C, where one client ran both channels in the same month, it went the other way: cold email replied at 9.1% per person, while the LinkedIn arm drew 5.4% in LinkedIn replies. Cold email also booked slightly more calls per 100 people there (0.4 against 0.3).

So I would not rank channels from reply rates across different clients. And if a vendor quotes a blended number from email plus LinkedIn, you cannot hold it against a cold-email-only benchmark.

One campaign, two readings of the same replies. C's email arm got 453 replies from 5,000 people and 15,000 emails.

Per person contacted, that is 9.1%, far above the 3.85% Smartlead needs for its top 5%. If some people replied more than once, fewer than 453 people wrote back, so read 9.1% as a ceiling.

Per email sent, which is Instantly's denominator, the same 453 replies are 3.0%, just under Instantly's 3.43% average. That is the denominator problem from the top of this page inside a single campaign.

Where we tracked both, fewer than 1 in 10 replies were interested. In A, 21 of 289 replies were interested (7.3%). In B, 26 of 309 became real conversations (8.4%). Measured against people contacted, those two LinkedIn-led campaigns had a positive reply rate of 1.75% and 1.2%.

In D, run as email outreach, 335 interested leads from more than 70,000 people emailed is under 0.5%.

So when someone asks what a good positive reply rate for B2B is, the honest answer starts with the channel and the list size: low single digits in our two LinkedIn-led campaigns, under half a percent in the 70,000-person email campaign.

Low rates can still pay at volume. That sub-0.5% interested rate in Campaign D still came to 141 calls in 60 days, from more than 70,000 people emailed with clean data. As I have put it before, "Volume negates luck." The rate tells you the cost of each call. The volume decides how many you get.

Four campaigns are a small set. I posted these numbers on LinkedIn between 2023 and 2025, and their channels and periods differ. Use them to see the shape of a funnel. Where your own numbers land depends on your list, your offer and the sender behind the email.

Common questions about cold email reply rates

What reply rate should a B2B cold email campaign hit in 2026? On a platform dashboard that counts all replies per email sent, 3% to 5% is normal and 5% or more is good, per Instantly's and Apollo's tiers. On a count of replies per person contacted, Smartlead's method, under 1% is typical, and about 1.4% or more puts you in its top quarter. Always compare within the same counting method.

Why do reply rate benchmarks vary so much between reports? Because they count different things. Some divide by emails sent, others by people contacted. Some count auto-replies and follow-ups, others do not. Each sample mixes different senders, from careful agencies to every account on a platform. The oldest widely quoted figure, 8.5%, is a 2019 link-building study, not sales email.

Does LinkedIn outreach get a different reply rate than cold email? Yes, but not always in LinkedIn's favor. Our two LinkedIn-led campaigns replied at 24% and 14%. When one client ran both in the same month, cold email replied at 9.1% per person against 5.4% on LinkedIn, and booked slightly more calls per 100 contacted (0.4 against 0.3). Compare channels on your own list, by calls booked.

What reply rate should I expect from a brand-new domain? Do not judge a reply rate until the domain has been set up and warmed properly and bounces are under 2%. Until then, low numbers mostly measure deliverability. After that, hold it to the same tiers as any other sender.

Check these six things before you compare your number

Before you decide your campaign is failing, or winning, write down an answer to each of these:

1. Which count does your tool use? All replies, or only real replies with auto-replies removed. Per email sent or per person contacted. Pick the matching benchmark column.
2. Is it cold email only? If LinkedIn or calls fed the same pipeline, separate the replies by channel before you compare.
3. How big was the list? A 40-person hand-picked list and a 4,000-person list are different tests.
4. How many follow-ups actually went out? Instantly's data puts 42% of replies on follow-ups. If the sequence stopped at one or two emails, you gave up a good part of that share.
5. Where is the bounce rate? Above 2%, fix delivery first. Most "bad reply rate" problems I see start in the data, which is why the list usually matters more than the email.
6. How many replies were positive? A 6% reply rate made of out-of-office notes and "remove me" is worse than 3% with a handful of real conversations. Then check whether each interested reply gets a same-day answer.

If all six check out and you are still below the typical line for your counting method, then look at the offer and the sender behind the email. Write the six answers next to the number in your next campaign review, and the benchmark finally has something fair to measure.