# Soft bounce: Definition and How It Works

Source: https://wraps.dev/glossary/soft-bounce

A soft bounce is a temporary delivery failure — a full mailbox, an unreachable server, an oversized message — reported with an SMTP 4xx code, meant to be retried rather than treated as permanent.

## What actually causes one

The common causes are a full mailbox, a receiving server that is temporarily down or overloaded, a message that exceeds the recipient's size limit, or a greylisting delay where the receiver deliberately asks the sender to try again shortly. None of these say anything about whether the address is real — they say the delivery attempt itself couldn't complete right now.

## The SMTP code family that signals it

A soft bounce corresponds to an SMTP 4xx response — 421, 450, 451, 452 are common — as opposed to the 5xx family that signals a permanent failure. A well-behaved sending system retries a 4xx with backoff for a bounded period before giving up, rather than either abandoning immediately or retrying forever.

## Retrying is expected, retrying forever is not

The correct response to a soft bounce is a retry, typically with exponential backoff over a window of hours to a couple of days, not an immediate resend and not an indefinite queue. AWS SES itself retries a temporarily undeliverable message automatically for a period before giving up and reporting a final bounce, so most SES senders don't need to build their own soft-bounce retry logic on top of it.

## Repeated soft bounces are a reputation signal too

A single soft bounce is noise. An address that soft-bounces on every single attempt over an extended period is functionally dead even though no individual attempt returned a permanent failure — a full mailbox that's never emptied, for instance, looks exactly like this. Reputation systems and mailbox providers both treat a persistently soft-bouncing address with growing suspicion the longer it goes on, even without a single hard bounce ever occurring.

## Why it matters

Treating a soft bounce like a hard bounce (suppressing the address immediately) throws away legitimate recipients over a transient condition, while treating a persistently soft-bouncing address like it's still healthy quietly degrades sender reputation over time without a single alarm-worthy event ever firing. The right handling sits between those two mistakes.

## How to check it

```
aws sesv2 get-suppressed-destination --email-address bounced@example.com
```

## Common questions

### Should I resend a message immediately after a soft bounce?

No. Retry with backoff over a period of hours to a couple of days, which is also what SES itself does automatically before reporting a final failure. Immediate resends waste attempts on a condition that usually needs time to clear.

### Does a soft bounce hurt my sender reputation the way a hard bounce does?

A single one, generally no. An address that soft-bounces on every attempt for weeks functions like a dead address for reputation purposes even though no individual attempt was a permanent failure, so persistence is what matters, not the type.

### Does AWS SES retry soft bounces for me automatically?

Yes, for a bounded period after a temporary failure, before it reports the message as a final bounce. Most senders on SES don't need to implement their own soft-bounce retry logic on top of that.