Soft Bounce a temporary delivery failure

The SMTP codes that produce a soft bounce

Soft bounces come back as 4xx codes. The 4 says "try again later." A good sending platform treats each of these differently because the reason matters for retry pacing.

• 421 · Service not available. The receiving server is busy, restarting, or actively throttling you. Retry with backoff. If 421s pile up from one ISP, you've probably been rate-limited and need to slow down.
• 450 · Mailbox temporarily unavailable. The address exists but the mailbox can't accept right now. Often locked or briefly busy. Retry in a few minutes.
• 451 · Local error in processing. Something went wrong on the receiver's side. Not your fault. Retry with backoff.
• 452 · Insufficient system storage. Mailbox or server is full. The mailbox owner usually clears space eventually. Retry over the next few hours.
• 421 with greylisting wording. Some receivers (Postfix postgrey is a classic) return 421 the first time a new sender hits them, expecting you to retry. Honest senders pass the test; spammers usually don't bother. This is by design.

Codes in the 5xx range (550, 551, 553, 554) are hard bounces, which means permanent. Those should be suppressed, not retried.

Why soft bounces still matter for reputation

A single soft bounce isn't a reputation event. A pattern of them is. If 8% of your sends to a given mailbox provider come back as 421s for a few days running, that provider has effectively told you to slow down. Ignoring that and retrying aggressively is the fastest way to turn soft bounces into hard ones and into reputation damage that lingers for weeks.

Industry rough target: soft bounce rate below 5% on a daily basis is normal. Above 10% sustained is a sign that something's wrong on your side, whether it's a content trigger, an authentication mismatch, or a bad list segment hitting a single ISP at once.

What to do when a soft bounce happens

In order of priority:

• Retry with exponential backoff. First retry in minutes, then an hour, then a few hours, then 24h apart. Don't hammer.
• Watch for ISP patterns. If Gmail starts returning 421s while Outlook stays clean, your problem is Gmail-specific. Slow down to Gmail and investigate before everything else.
• Promote to hard after enough failures. Most platforms reclassify a soft bounce as hard after the address has soft-bounced 5–7 sends in a row over several days. That address is effectively dead even if the SMTP code keeps saying "later."
• Investigate the cause if a 4xx spikes. Sudden surge of 452s could mean a single domain's mail store is broken. Sudden surge of 421s could mean you're being rate-limited.

Hard bounce vs soft bounce

Same surface (the address didn't get the mail), totally different mechanics underneath:

• Hard bounce (5xx codes): permanent. Suppress immediately. Never retry.
• Soft bounce (4xx codes): temporary. Retry with backoff. Reclassify and suppress only after repeated failures.

How sendmsg.io handles soft bounces

Every SMTP response on our platform gets parsed in real time by the Cortex engine. Soft bounces go straight into a per-recipient retry queue with exponential backoff (minutes, then an hour, then a few hours, then 24h). If an address soft-bounces 5 sends across that window, it gets reclassified and added to the suppression list pending manual override.

Per-ISP soft bounce rates also feed into the domain reputation temperature. If we start seeing 421s climb from one specific mailbox provider, the engine applies graduated throttle to that destination before the provider has to escalate to outright blocks. That's the whole point of reputation management: respond to the warning, don't wait for the punishment.