Deliverability

Hard Bounce vs Soft Bounce: What to Do About Each

Hard bounces are permanent, soft bounces are temporary. Learn how to read the codes and the exact action each one needs.

07 Jun 20266 min read
Hard Bounce vs Soft Bounce: What to Do About Each

A bounce is a delivery failure reported by the receiving server. There are two families, and they need different treatment.

Hard bounce: permanent

The address does not exist, the domain is invalid, or the server permanently refuses mail. These usually come with 5xx codes.

Action: remove the address immediately. Retrying only hurts your reputation.

Soft bounce: temporary

The mailbox is full, the server is down, or the message is too large. These come with 4xx codes.

Action: retry for a short period. If the same address soft bounces across several campaigns, treat it as dead and remove it.

Blocks are a third case

Some bounces say the message was blocked because of reputation or content. These are about you, not the address. Check authentication, content and complaint rates.

A simple rule set

  • 1 hard bounce: remove
  • 3 soft bounces in a row: remove
  • Block messages: investigate your sending, do not remove the contact

Prevent them

Verification catches most future hard bounces before you send. Use the email checker or the bulk verifier.

Reading the bounce code

Bounce messages carry a status code that tells you what kind of failure it was. The first digit is the key: 5.x.x is permanent, 4.x.x is temporary.

  • 5.1.1 means the recipient does not exist: a hard bounce, remove the address.
  • 4.2.2 means the mailbox is full: a soft bounce, retry later and stop after a few attempts.
  • 5.7.1 is usually a policy block, often about your sending reputation or authentication, not about the address.

Our error code library explains each code and what to do about it.

A practical suppression rule

  • Remove hard bounces immediately and permanently.
  • For soft bounces, retry up to three times over a few days, then suppress the address if it keeps failing.
  • Treat a policy block (5.7.x) as a sender problem. Do not delete the contact. Fix authentication or reputation, then resend.

Why the distinction pays off

If you treat every bounce as final, you lose customers whose mailbox was only full for a weekend. If you treat every bounce as temporary, you keep sending to dead addresses and damage your reputation. Reading the code is the cheap way to get both right.

Verifying before you send removes most hard bounces before they happen. Start with the bulk verifier.

A retry rule for soft bounces

Do not give up on the first temporary failure. A common rule is to retry a soft bounce on the next one or two sends. If the same address soft-bounces three times in a row, treat it as permanent and remove it. Most sending tools do this on their own; check the setting.

Special cases

  • Mailbox full: retry once, then drop it if it stays full after a week or two.
  • Message too large: shrink images and attachments before the next send.
  • Blocked by policy or reputation: this is not about the address. Check your records and your reputation before you retry.
  • Auto-reply "no longer works here": treat it like a hard bounce and remove the address.

Keep the counts separate

Track hard and soft bounces as two numbers. A rising hard-bounce count points to list quality. A rising soft-bounce count often points to server limits, message size or reputation.

A short routine

After each send, export the bounce list, remove hard bounces the same day, mark soft bounces with the date and review them before the next send. It takes ten minutes and keeps the list healthy.

Counting bounces correctly

Report hard and soft bounces as two separate numbers. A combined rate hides the problem: a list with many soft bounces during a holiday week looks bad but recovers, while a rising hard bounce rate means the list itself is decaying.

What the numbers should look like

Bounce rates are only useful when you read them in the right unit. Report hard bounces as a percentage of delivered attempts for each send, not as a monthly total, because one bad import can hide inside a good month. For a list that has been cleaned and is mailed regularly, a hard bounce rate that stays well under one percent is realistic. A single send above two percent is a warning that something entered the list that should not have. Soft bounces fluctuate more, and a sudden rise usually points to a temporary problem on your side or the receiver's, such as a throttling response or a full mailbox wave after a holiday.

A practical retry schedule for soft bounces

Do not retry soft bounces forever and do not give up after the first one. A sensible rule is to let your sending system retry within the same delivery window, then keep the address in the next two sends. If it soft-bounces in three consecutive campaigns, move it to the suppression list and check it again with a verifier after a month. If the verifier says the mailbox exists and accepts mail, you can release it; if the answer is invalid or unknown, leave it suppressed.

Mapping codes to decisions

You do not need to memorise every SMTP code. Learn the pattern. Codes starting with 5 mean the receiving server has refused the message for good, and codes starting with 4 mean it asked you to try again later. Inside those families the text tells you more: unknown user, mailbox disabled and domain not found are address problems you fix by removing the contact. Message too large, content rejected or blocked by policy are problems with the message or with your reputation, and removing the address will not help. Keeping a short table of these readings next to your reporting makes the weekly review faster.

Do not let a bounce delete a good customer

Automatic removal is convenient but blunt. If your system removes any address on the first hard bounce, a typo in an old record can remove a customer who still exists under a corrected address. Before deleting, keep the record in an archive state with the reason and date. If the customer later contacts support, you can restore the correct address instead of creating a duplicate. For transactional senders, such as order confirmations, a hard bounce should also raise a flag in your own system so someone can ask the customer for a working address.

Prevent the next one

The cheapest bounce is the one that never happens. Check addresses at the point of capture, check imported lists before the first send, and re-check any address that has been quiet for months before a large campaign. Those three habits cover almost every source of hard bounces, and they leave soft bounces as the small, mostly harmless noise they should be.

Mailbite Editorial TeamPractical guides on email verification and deliverability.