Cold Email Benchmarks 2026: Reply Rates by Industry and What "Good" Looks Like

The headline numbers, and why they disagree

Search "cold email reply rate benchmarks" and you will find numbers ranging from 0.45% to 8%. That spread is not noise. Each report is measuring something different, and knowing which number applies to you matters more than the number itself.

Source	Dataset	Average reply rate
Instantly benchmark report	Billions of interactions, Jan-Dec 2025	3.43% of delivered
Apollo (compiled)	Multiple 2025-2026 datasets	~3.1%, well-run range 3-5%
Belkins study	7.5M emails, 2025 campaigns	0.45% of total sent

Three things drive the gap. First, the denominator. Instantly counts replies against delivered emails. Belkins counts against every email sent, including ones that bounced or never landed. Second, what counts as a reply. Some datasets include out-of-office messages and unsubscribe requests, which inflate the number without meaning anything for pipeline. Third, sample bias. Sending-platform data over-represents senders who already invested in warmed inboxes and proper infrastructure. The strict studies include everyone, including the teams blasting unverified lists from a single domain.

The practical read: if your reporting counts total replies against delivered emails, which is what most sending tools show you, then 3% to 5% is the honest 2026 average for a properly built campaign. If you are below 2%, something structural is broken. If someone promises you 15%, ask which denominator they are using.

Reply rates by industry

Industry changes the baseline more than most copy decisions do. The ranges below compile 2026 benchmark data from Cleverly's industry breakdown, with the caveat that these are compilations of platform data, not audited studies. Treat them as directional.

Target industry	Typical reply rate
Legal services	~8-10%
EdTech / e-learning	~5-7%
Real estate	~5-7%
Healthcare and MedTech	~4-6%
Consulting	~4-6%
Financial services	~3-4%
IT services and MSPs	~3-4%
SaaS / software	~2-4%
Consumer goods	~1-3%

The pattern behind the table is inbox saturation. Legal, real estate, and healthcare buyers receive comparatively little cold email, so a relevant message stands out. SaaS founders and RevOps leaders receive the most cold volume of anyone on the internet, which is why SaaS-to-SaaS outreach sits near the bottom despite having the most sophisticated senders.

Belkins' stricter dataset adds two cuts worth knowing. By seniority, founders and owners replied at 0.57%, C-level at 0.42%, and VPs at 0.32%. By company size, companies with 10 or fewer employees replied at 0.72% versus 0.22% at companies with 10,000 or more. Small company founders reply roughly three times as often as enterprise contacts. If your ICP allows it, aim low in the org chart of small companies rather than high in the org chart of large ones.

What good looks like at each tier

Apollo's 2026 compilation and Instantly's quartile data agree on the shape of the performance curve, so here is the tier map I would actually use.

Total reply rate	What it means
Below 2%	Deliverability or list problem. Fix infrastructure before touching copy.
3-5%	Normal for a well-run 2026 campaign. Solid ICP, clean list.
5-8%	Top quartile. Instantly puts the top-quartile cutoff at 5.5%.
10%+	Top tier (Instantly's cutoff: 10.7%). Almost always signal-triggered outreach to narrow segments.

One correction to how most teams read these tables: total reply rate is a diagnostic metric, not a success metric. It includes out-of-office autoresponders, polite declines, and "take me off your list." The number that predicts pipeline is positive reply rate. Practical 2026 benchmarks there, per Apollo: 2-3% is minimum viable, 4-6% is the target, 8%+ is a stretch. A campaign at 6% total replies where half are negative is worse than a campaign at 4% where most replies are questions about your offer.

The numbers behind the numbers

A few second-order stats from the same reports explain more about performance than the headline averages do.

Small campaigns beat large ones by almost 3x. Campaigns targeting 50 or fewer recipients average a 5.8% reply rate. Scale the same motion to 1,000+ recipients and the average drops to 2.1%. Nothing about the sending changed. What changed is that a 50-person list can be genuinely relevant to every person on it, and a 1,000-person list cannot.

Step one carries the campaign. 58% of all replies come from the first email in the sequence, per Instantly's 2026 report. Follow-ups matter, and the data supports 4 to 7 touchpoints before returns diminish, but no follow-up cadence rescues a first email that did not land.

Shorter wins. The same report puts the ideal first touch under 80 words. Multiple 2025-2026 datasets show 50 to 125 word emails outperforming longer formats by roughly 50% on replies.

Bounce rate has a hard ceiling. Keep bounces under 2%, ideally well under. Above that, mailbox providers start treating your domain as a spam source, and every subsequent send performs worse regardless of how good the copy is. This is a list verification discipline, not a sending trick.

Why open rate is not in this article

The average cold email open rate in 2026 is around 27.7% per industry compilations, and I would not plan anything around it. Apple Mail accounts for roughly half of recorded opens and preloads tracking pixels automatically, which means a large share of "opens" are machines, not people. Worse, open tracking pixels themselves hurt deliverability with some providers. Most serious senders in 2026 have turned open tracking off entirely and measure replies, positive replies, and meetings. I would suggest the same.

How campaigns actually beat the benchmark

Look back at the spread in every table above. Legal outreach at 8-10% versus SaaS at 2-4%. Fifty-person lists at 5.8% versus thousand-person lists at 2.1%. Founders at small companies replying at three times the rate of enterprise VPs. The gap between average and top quartile is never explained by subject lines or send times. It is explained by who is on the list and whether the message is about their actual situation.

That is the uncomfortable part of benchmark data. Teams stuck at 1.5% usually go looking for better copy. But the campaigns clearing 5.5% are not winning on wording. They are winning on three inputs that happen before anyone writes a sentence.

List quality. Every contact verified before sending, so bounces stay under 2% and the domain reputation compounds instead of eroding. Every contact actually matching the ICP, not just matching the Apollo filter that approximates the ICP. The difference between those two lists is invisible in the CSV and enormous in the reply data.

Segment size. The 5.8% versus 2.1% list-size stat is really a relevance stat. Small segments force a specific message. If a campaign cannot say something true and specific about everyone on the list, the list is too broad. Splitting one 800-person campaign into eight segments of 100, each with its own angle, is tedious. It is also roughly what the data says the top quartile does.

A reason to write. The 10%+ tier is dominated by signal-triggered outreach: a fundraise, a hiring spike, a product launch, a leadership change. The email exists because something happened, and the recipient can tell. Volume-first outreach can be dressed up to imitate this. Recipients in 2026 have seen enough cold email to know the difference.

What I would do with these numbers

Set your baseline expectation from the industry table, not from a generic average. Measure positive replies per 1,000 delivered as your primary metric. If total replies sit below 2%, diagnose infrastructure and list quality first and do not touch the copy until those are ruled out. If you are at 3-5% and want more, the lever is narrower segments and better reasons to write, not a fourth follow-up.

And hold the benchmarks loosely. Every number in this article is a median over thousands of senders with different lists, offers, and infrastructure. Your campaign is a sample size of one. The benchmark tells you whether something is structurally wrong. It does not tell you what your specific ICP will do, and anyone quoting you a guaranteed reply rate before seeing your list is quoting you a number from someone else's campaign.