What bounce rate is acceptable for cold email?
Keep hard bounces under 2% per mailbox. What the 2026 data shows, hard vs soft bounces, and what to do when you pass 4%.
Keep hard bounces under 2% per mailbox, measured over the last two weeks. Between 2% and 4%, stop adding new leads and re-verify the list. Above 4%, pause the mailbox until you know why. And whatever the overall number, any campaign that bounces 10 out of its first 100 emails has a list problem.
One thing to know before you trust any of these numbers: no mailbox provider publishes a bounce threshold. Google publishes a spam-rate limit, not a bounce limit. The 2% line comes from people who run sending platforms and watch what happens to mailboxes that cross it. It is a good line. It is just not a rule written by Gmail.
What the data actually shows
The most useful dataset right now is Smartlead's State of Cold Email report, built from more than 850 million emails sent through its platform in the first half of 2026, counting only senders with at least 500 emails in the period (Smartlead).
- median bounce rate across cold email senders
- 1.54%
- the line the best quarter of senders stay under
- 0.73%
- the bounce rate a quarter of senders hit or exceed
- 3.21%
So a typical sender bounces about 1 or 2 emails per hundred, the careful ones under 1, and a quarter of all senders are running above 3%. That last group is where domains get burned. If your number is 3% and you think "that's basically fine", you are in the worst quarter of the market.
It's vendor data from one platform, so treat it as a benchmark, not a law. But it lines up with what you see in practice: the people who verify every list sit under 1%, the people who scrape and send sit at 3–6%.
Hard bounce, soft bounce: which one counts
Not every bounce is the same, and mixing them up leads to wrong decisions in both directions.
The codes in the bounce message tell you which kind you have. The standard (RFC 3463) splits them by the first digit: 5.x.x is a permanent failure, "not likely to be resolved by resending", and 4.x.x is a transient one where "sending in the future may be successful".
| Bounce | Typical code | What it means | What to do |
|---|---|---|---|
| Hard | 5.1.1 | The mailbox does not exist | Remove the address today and never send to it again |
| Hard | 5.7.x | Rejected by policy: authentication, reputation, block | Stop the mailbox and check SPF, DKIM, DMARC and blocklists |
| Soft | 4.2.2 | Mailbox full | Retry later; drop it after two or three failures |
| Soft | 4.x.x | Temporary deferral, often throttling | Lower volume on that mailbox |
The 2% target is about hard bounces of the 5.1.1 kind: addresses that don't exist. They are the clearest signal a provider gets that you are mailing a list you didn't build carefully.
The 5.7.x rejections are a different problem. Your list may be clean and the provider still refuses you because of authentication or reputation. Re-verifying addresses won't help there.
Soft bounces matter in their own way. Google's own guidance for senders says: "If messages start bouncing or start being deferred, reduce the sending volume until the SMTP error rate decreases" (Gmail sender guidelines). A wave of 4.x.x deferrals is the provider telling you to slow down, before it starts saying no.
Why bounces hurt more in cold email
A newsletter bounce costs one subscriber. A cold email bounce costs reputation on a mailbox that has no history to absorb it.
Cold mailboxes send to people who have never heard of you, so they get few replies and few opens compared with normal mail. They live on thin engagement. Every dead address is evidence that you bought or scraped the list, and Google tells senders plainly: "Don't purchase email addresses from other companies." The same page tells senders to "Automatically unsubscribe recipients who have multiple bounced messages".
The other reason is decay. B2B addresses rot because people change jobs. A list that verified clean in March will bounce in September, even if you never touched it. That is why a one-time verification is not enough for a list you work for months.
How to get under 2% and stay there
Keeping cold email bounces under 2%
- Verify right before the sendNot when you build the list. Run every address through a verifier within a week of the first email.
- Treat catch-all addresses separatelyCatch-all domains accept everything at check time and bounce later. Send to them in small batches from your most established mailbox, or skip them.
- Check the person, not just the addressIf someone left the company, their address will start bouncing soon. A quick look at their current job saves the send.
- Watch the first 100 per campaignMeasure bounces per campaign and per mailbox. Ten hard bounces in the first hundred means stop and re-verify the rest.
- Remove hard bounces automaticallyA 5.1.1 address must never get a follow-up. Follow-ups to dead addresses double the damage.
A few things that sound reasonable and make it worse:
Do
- Verify within a week of sending
- Track bounces per mailbox
- Pause one mailbox, keep the rest running
- Remove the address after the first hard bounce
Don't
- Guess addresses from first.last@ patterns
- Send to a list verified months ago
- Average bounces across all mailboxes and call it fine
- Keep follow-ups going to addresses that bounced
The averaging point matters more than it looks. If you run ten mailboxes and one of them is at 7% while the others sit at 1%, your overall number is around 1.6% and looks healthy. That one mailbox is still getting burned, and its domain may be shared with the others.
What to do when you are already above 4%
- Pause the affected mailbox. Not the whole operation. Find which campaign and which list the bounces came from.
- Read the codes. Mostly
5.1.1means a list problem. Mostly5.7.xmeans an authentication or reputation problem, and the fix is in DNS and blocklists, not the list. - Re-verify the remaining leads of that campaign before anything else goes out to them.
- Restart at half the volume you were sending, and only move back up after a week of bounces under 2% and replies coming in.
- Check spam placement with a few seed mailboxes of your own on Gmail and Outlook, and watch Google Postmaster Tools for complaints.
If replies dropped to zero at the same time as bounces went up, assume you are landing in spam already. Recovery then takes weeks, and for a secondary sending domain it is often cheaper to warm up a new one. We covered how long that takes in How long to warm up a new email domain.
Where the bounces really start
Most bounce problems are decided before the email is written: in how the list was built and how long it sat. That's the part worth automating. Nitrino finds people in its own base of contacts with verified email addresses, and before every first letter it reads the person's current LinkedIn profile, so someone who has changed jobs gets updated instead of mailed at a dead address. It also paces each mailbox by its health rather than by a fixed daily number. If that is the problem you have, here is how Nitrino works.
Whatever tool you use, the rule is the same: measure hard bounces per mailbox, keep them under 2%, and stop at 4%.
