How a hard bounce happens

During an SMTP conversation, the sending server announces each recipient with a RCPT TO command and the receiving server replies with a three-digit code. Under RFC 5321, codes starting with 2 mean success, 4 means a temporary failure, and 5 means a permanent failure. A hard bounce is a 5xx reply that tells the sender this message will not be delivered and should not be retried as-is.

Most servers also return an enhanced status code defined in RFC 3463, written as class.subject.detail. The class digit mirrors the basic reply (5 is permanent), and the subject and detail explain why. If the rejection happens after your mail server has already accepted the message (for example, at a relay further down the chain), the failure comes back as a bounce message in the Delivery Status Notification format from RFC 3464, carrying the same codes.

Common hard bounce codes

  • 550 5.1.1 (bad destination mailbox): the mailbox doesn’t exist. This is the classic hard bounce, typically caused by a typo, a deleted account, or a made-up address.
  • 550 5.1.2 or 5.1.10 (bad destination system): the domain doesn’t exist or publishes a null MX record declaring that it accepts no mail.
  • 550 5.2.1 (mailbox disabled): the account exists but has been deactivated and is not accepting messages.
  • 550 5.7.1 (delivery not authorized): the server refused the message on policy grounds. This is a permanent rejection of that send, but it is usually about the sender’s reputation, authentication, or content rather than the recipient address.

The exact text after the code varies by provider, so rely on the codes rather than the human-readable message. The full SMTP bounce code reference lists what each common code means and whether to suppress the address.

Why hard bounces matter

Mailbox providers watch how often a sender hits addresses that don’t exist. A high rate suggests purchased, scraped, or stale data, and it feeds directly into sender reputation. That reputation then affects inbox placement for every recipient, including the engaged ones.

Sending platforms enforce this too. Amazon SES, for example, recommends keeping bounce rate below 2%, places accounts under review at 5%, and may pause sending at 10%, counting only hard bounces toward that rate. Many email service providers apply similar internal thresholds. See what bounce rate is safe for more context.

There is also a direct cost: every hard bounce is a send you paid for that reached no one, and old invalid addresses are a known source of recycled spam traps.

How to handle hard bounces

  1. Suppress immediately. Add any address that returns 5.1.x or 5.2.1 to a permanent suppression list after the first failure. Don’t wait for a second attempt.
  2. Separate address failures from sender blocks. If a large share of bounces in a send are 5.7.x policy rejections from one provider, the problem is your reputation or authentication (check SPF, DKIM and DMARC alignment), not your list. Suppressing those addresses hides the real issue.
  3. Make suppression global. A hard-bounced address should be excluded from every stream and tool you send from, including your CRM, your marketing platform, and your transactional mail.
  4. Verify before sending. Pre-send verification catches most future hard bounces. An address a verifier marks undeliverable would have hard-bounced if you had mailed it.
  5. Re-verify aging lists. Lists decay as people change jobs, so verify any segment you haven’t mailed in a few months before sending to it again.

Where RocketVerifier helps

RocketVerifier checks each address over SMTP without sending a message, so addresses that would hard-bounce come back as undeliverable with a specific sub_status such as mailbox_not_found, account_disabled, or does_not_accept_mail. Policy rejections are reported as unknown with sub_status blocked, so a sender-side block is never mistaken for a dead mailbox. You can clean a full list with bulk email verification or check a single address with the free email verifier.