HomeResourcesGuides
Technical guide

Hard bounce vs soft bounce: differences and what to do with each

Not all bounces mean an address is invalid. Understanding the cause helps decide when to wait, investigate, or suppress a recipient.

What bounces are and what to do with them

When an application sends transactional email, it is not enough to know the provider accepted the request. Delays, rejections, and subsequent decisions matter too.

A bounce is a rejection or return reported during delivery. It may relate to the address, content, recipient policy, authentication, or a temporary issue.

Classification requires context

Interpretation depends on the provider, SMTP code, and available events. The same diagnostic should not always lead to the same action for a recipient.

Potentially permanent error

What is a hard bounce

A hard bounce usually needs attention, but it does not prove on its own that the address is invalid. The specific cause should decide whether to suppress the recipient or review the send.

Common causes

  • Non-existent address.
  • Non-existent domain, a domain configured not to receive email, or one without a resolvable SMTP destination.
  • Message rejected due to content or size.
  • Recipient policy.
  • An authentication or reputation issue.

What to do

  • Identify whether the diagnostic is related to the address.
  • Suppress the recipient only if the address does not exist or cannot receive email.
  • Avoid further automatic sends to a suppressed address.
  • Record the reason and allow correction when appropriate.
Temporary error

What is a soft bounce

A soft bounce usually means delivery cannot complete at that moment. The address may still be valid and the provider may attempt delivery again later.

Common causes

  • Full mailbox.
  • Temporarily unavailable receiving server.
  • Timeout or greylisting.
  • Temporary receiving limit.
  • Temporary DNS error.
  • Throttling or a temporary reputation issue.

What to do

  • Record the event and its diagnostic.
  • Wait for the send already in progress to evolve.
  • Do not immediately block the recipient.
  • Do not create another message while the provider may still retry.
  • Review the final failure if it arrives.

Who should perform the retry

For a temporary error, first check how your provider behaves. Many platforms automatically retry delivery for a set period. The application should record delay and final failure events, but not create a new send while the previous one may still be in progress. Reserve application retries for requests the provider did not accept, failures before they entered its system, or an explicit business decision.

Operational decision

Hard bounce vs soft bounce

AspectHard bounceSoft bounce
NatureIt may be permanent, but needs a diagnosis.Usually temporary.
ExamplesNon-existent address, rejected content, or policy.Full mailbox, timeout, or greylisting.
First actionClassify the cause before suppressing.Wait for provider retries.
Risk if ignoredIncorrect suppression or continued sends to a non-existent address.Duplicate messages or block a valid address.

Decisions by cause

Non-existent address

Suppress the recipient and request correction.

Full mailbox or temporary error

Wait for provider retries.

Rejected content

Review the message.

Policy or reputation

Review authentication and reputation.

Message too large

Correct the content.

States and traceability

Minimum recommended states

Keep the model simple. Store bounce type and cause as event attributes, not as a separate state for every diagnostic.

pending

The message is prepared or queued.

accepted/sent

The provider accepted the message; it does not imply delivery.

delivered

The receiving server accepted the message; it does not imply reading or inbox placement.

soft_bounced

The provider reported a temporary bounce.

hard_bounced

There is a permanent bounce; review its cause before suppressing.

rejected

The provider or receiver rejected the message.

failed

The send could not complete.

suppressed

No further automatic sends should occur.

Information worth keeping

  • Internal identifier.
  • Provider identifier.
  • Event date.
  • Provider classification.
  • SMTP or enhanced code.
  • Server text response.
  • Attempt number, when available.
Recognizable situations

How bounces become an incident

An invoice that never arrives

If a bounce does not update the customer record, later invoices will keep going to an address already known not to work.

A blocked password recovery

The system may generate the email correctly, but the user still cannot access their account and will blame the application.

A difficult incident to investigate

Without code, reason, and identifiers, support and development must search several systems to explain what happened.

Common mistakes

What to avoid

Treating all bounces the same

It may suppress valid addresses or ignore a real issue.

Classify the cause before deciding.

Resending while the provider retries

It may create duplicate messages.

Wait for events from the send already in progress.

Blocking after a temporary failure

A valid address may stop receiving critical messages.

Reserve suppression for address diagnostics.

Suppressing for content or reputation

The issue will continue to affect other recipients.

Review the message, authentication, or reputation.

Not retaining the original diagnostic

The incident will be hard to explain and investigate.

Keep codes, response, and identifiers.

Quick check

Before closing a bounce, confirm that:

  • The application distinguishes acceptance from delivery.
  • Events update the message state.
  • The original diagnostic is retained.
  • Address errors trigger suppression.
  • Content or reputation errors are investigated.
  • Provider retries are not duplicated by the application.
  • There is a clear limit for any application retry.

Relation to pre-send validation

Pre-send validation can reduce some hard bounces by detecting format errors, non-existent domains, or risk signals before sending.

It does not replace bounce handling: validation acts before sending and delivery events allow a reaction afterwards.

Mailiqo

Manage sends and their errors from one place

Mailiqo lets you send transactional emails through API or SMTP and view their traceability, errors, and events from the same platform.

Learn about Mailiqo
Mailiqo Labs

Put the guide into practice

The tools let you check SPF, DMARC, headers, bounces and HTML with your own data, without registering.