Benchmarks are useful for one thing: telling you which part of your campaign is broken. They are close to useless as targets, because a "good" reply rate in one market is a disaster in another.
Here is a set of realistic ranges, and more importantly, how to read them together.
The numbers
| Metric | Healthy | Investigate | Broken |
|---|---|---|---|
| Reply rate | 3 to 8% | 1 to 3% | Under 1% |
| Positive reply rate | 1 to 3% | 0.3 to 1% | Under 0.3% |
| Bounce rate | Under 1% | 1 to 3% | Above 5% |
| Complaint rate | Under 0.1% | 0.1 to 0.3% | Above 0.3% |
| Unsubscribe rate | Under 1% | 1 to 3% | Above 3% |
Two notes before you use these.
First, they assume cold B2B outreach to a verified list from a warmed domain. Warm audiences, existing customers and inbound follow-ups behave completely differently.
Second, market matters more than any benchmark. Selling something unusual to a well-defined niche can produce reply rates that look implausible. Selling a crowded category to a broad list can produce 1% with excellent copy. Your own previous campaign is a better comparison than anyone's published average.
Why open rate is gone
Apple Mail Privacy Protection pre-fetches images in messages, including the tracking pixel that measures opens, whether or not the recipient ever looks at the message. A large share of email now arrives through clients that do this.
The consequence is that a reported open rate is some unknown mixture of real opens and automatic pre-fetches, and the mixture changes with your audience's device preferences. It is not that opens got less accurate; it is that they no longer measure a consistent thing.
Three practical implications:
- Do not report open rate to anyone who will make a decision on it.
- Do not A/B test subject lines on opens. Test them on replies, which requires more volume and more patience, and is the only version that means anything.
- Be suspicious of any tool or agency quoting a 70% open rate as evidence of quality.
Reading the shape
Individual numbers are less informative than their relationships. Four common patterns:
Low bounces, low replies
Deliverable, ignored. Almost always targeting or offer, not copy. You reached the people; they did not care.
Fix the list before you touch the writing. The question is whether these people plausibly have the problem you solve, right now.
High bounces, low replies
List quality. The bounces are the visible damage; the invisible damage is what the bounce rate did to your placement, which then suppressed everything else.
Verify the list, purge the invalids, and re-verify anything older than a few months. Then send again from a rested domain.
Decent replies, no positive replies
You are reaching people and provoking a response, but the response is "not me" or "not now". Usually a targeting near-miss: the right industry, the wrong role or the wrong stage.
This is a good problem. It means the messaging works and the aim is slightly off.
Everything near zero
Deliverability. Before changing anything else, check where your mail is actually landing: send to seed addresses across Gmail, Outlook and a business domain and look.
A campaign in the spam folder produces exactly this shape, and every hour spent rewriting subject lines while it persists is wasted. Start with the deliverability checklist.
The maths of a meeting
Rough arithmetic that makes planning honest:
- 1,000 emails sent
- 980 delivered (2% bounce)
- 50 replies (5% of delivered)
- 20 positive (2% of delivered)
- 10 calls booked (half the positives)
- 2 to 3 opportunities
So roughly 100 emails per booked call at healthy rates. If you need ten meetings a month, that is 1,000 emails a month, which is about 35 a day, which is one warmed mailbox.
Two things fall out of that.
You need less volume than you think if the list is good. People reach for more sending because it is the easiest lever, not because it is the effective one.
Halving the list and doubling the research usually wins. A tightly targeted list of 50 businesses you genuinely understand will beat 1,000 names from a filter, and it will cost you less reputation to find out.
What to actually watch
Weekly, per campaign: reply rate, positive reply rate, bounce rate.
Weekly, per mailbox: bounce rate and complaint rate. An aggregate number hides the one mailbox quietly failing, and by the time the campaign average moves, that mailbox has been damaging your domain for a fortnight.
Monthly: domain reputation in Google Postmaster Tools, and a manual placement test to seed addresses. Placement is the only measurement that answers the question you actually care about.
The one benchmark that matters
Your last campaign.
Published averages are aggregated across markets, list qualities and offers that have nothing to do with yours. The useful comparison is whether this campaign beat the one before it, with volume held constant so you know it was the messaging and not the sending.
Hold volume flat. Change one thing. Measure replies. That loop, run four times, will teach you more about your market than any benchmark table, including this one.