If you send email at any volume, you will see bounces. Most senders lump them all together and either ignore them or hard-suppress every bouncing address. That is the wrong default in both directions. Hard bounces and soft bounces mean very different things, and the right handling depends on which one you are looking at.
Hard bounce
A permanent delivery failure. The receiving server is telling you, plainly, that this email cannot be delivered now or ever. The SMTP response code is in the 5xx range, most often 550.
What triggers a hard bounce
- The mailbox does not exist (
550 No such user). - The domain does not exist (
NXDOMAIN). - The mailbox has been closed by the user or admin.
- You have been explicitly blocked by the recipient's server.
How to handle hard bounces
Add the address to your suppression list immediately and permanently. Never try to send again, even if someone re-imports it later. Most ESPs do this automatically, but check that your import workflow respects suppression lists.
Soft bounce
A temporary delivery failure. The receiving server cannot accept the message right now, but might be able to later. SMTP response code is in the 4xx range, most often 421 or 451.
What triggers a soft bounce
- The mailbox is full.
- The receiving server is temporarily down.
- The message is too large (often above 10 MB).
- The sender is being rate-limited.
- Greylisting (the server is asking you to retry in a few minutes).
- Out-of-office responders interpreting your message as spam.
How to handle soft bounces
Retry. Most ESPs retry automatically for 24 to 72 hours with exponential backoff. Two patterns to watch for:
- One soft bounce, then delivers. Normal. No action needed.
- Soft bounces every send for weeks. Treat the address like a hard bounce. The mailbox is probably full and abandoned, or the receiving server has quietly blacklisted you.
Most ESPs convert chronic soft bouncers to suppressed automatically after 3 to 5 consecutive failures. If yours does not, build the logic yourself.
The grey area: catch-all and greylisting
Two responses sit in the middle:
Catch-all happens at verification time. The receiving server accepts every address at SMTP. It is not a bounce, it is an ambiguous green light. Treat catch-all addresses as risky, not safe.
Greylisting is a soft bounce by design. The receiving server is asking you to retry. ESPs handle this automatically. If you see persistent greylisting from one domain, talk to its admin or just accept the slightly higher latency.
Why the distinction matters
Treating every soft bounce as a hard bounce shrinks your list unnecessarily and removes engaged subscribers whose mailboxes were briefly full. Treating every hard bounce as a soft bounce keeps invalid addresses on the list forever, which destroys your bounce rate.
The right move is to do both: hard-bounce hard, soft-bounce gracefully, and verify the entire list quarterly to catch the addresses that should have already been removed.
FAQ
How does email verification handle this?
A verifier classifies a hard-bounce-likely address as invalid before you ever send to it. Soft-bounce-likely (greylisting, temporary issues) gets the risky label. You can pre-segment your list this way.
What is a normal hard bounce percentage?
Below 0.5% for healthy lists. Above 2% means your list is decaying or was bought.
Can I reach a hard-bounced address again?
If the recipient explicitly tells you they have re-activated the mailbox, you can manually remove from suppression. Otherwise, no.
Prevent bounces in the first place
The cheapest bounce is the one you never deliver. Pre-verify your list and the bounce rate drops to under 1% on the first send.