Email bounce rate: what is acceptable, and how to fix a spike

Bounce rate is the first deliverability number a receiver can measure about you and the easiest one to keep clean. It is also the one that spikes suddenly — one bad list, one catch-all domain, one stale export — and takes weeks to live down. Here is what the numbers mean and what to do when they move.

Hard and soft bounces

• Hard bounce (5xx): the address does not exist, the domain has no mail server, or the recipient's server has permanently rejected you. Remove the address immediately and never send to it again.
• Soft bounce (4xx): temporary — mailbox full, server busy, greylisting, or rate limiting. Retry a few times over a day or two; if it keeps failing, treat it as hard.

The dangerous middle case is a 4xx that is really a reputation signal. Gmail's 421-4.7.28 (“unusual rate of unsolicited mail”) looks like a soft bounce but means you are being throttled. If soft bounces cluster on one receiver, stop and read the response text.

What is acceptable

Hard bounce rate	What it means
Under 1%	Verified list, healthy sending. Where you want to live.
1–2%	Acceptable, but the list has drift. Re-verify anything older than 60 days.
2–5%	Receivers are noticing. Pause, verify, resume at lower volume.
Over 5%	You are being throttled or blocked. Stop the campaign now.

Measure per mailbox and per campaign, over the last 200–500 sends, not lifetime. A lifetime average hides a spike until it is too late.

Why spikes happen

• An unverified or old list. B2B addresses decay at roughly 2–3% a month as people change jobs.
• Catch-all domains. The domain accepts every address at SMTP time and bounces later, so verification says “valid” and the send says “no such user”.
• Role addresses and guessed patterns. firstname.lastname@ guesses from a data tool are a coin flip.
• A tool that retries hard bounces. Some sequencers retry everything; each retry to a dead address is another bounce.

Six-step recovery

1. Pause every sequence on the affected mailbox. Not the campaign — the mailbox. Reputation is per mailbox and per domain.
2. Export the bounces and classify them. Hard, soft, throttling. Delete hard bounces from every list and add them to global suppression.
3. Re-verify the remaining list with a service that detects catch-all domains. Segment catch-alls and send to them at a fraction of normal volume, or not at all.
4. Check for throttling. If Gmail or Outlook responses mention rate or reputation, wait 48–72 hours before resuming and halve the daily cap when you do.
5. Resume at half volume and let the cap rise only while the bounce rate stays under 1% for a few days.
6. Fix the source. If the list came from an export, set an expiry on lists (60 days) and verify at import time, every time.

Prevention

Verification at import, a hard bounce threshold per mailbox, and a dispatcher that stops when the threshold trips. Quotarider's guardrail holds any mailbox whose bounce rate crosses 2% on the last 200 sends until the list is re-verified, and the daily cap cannot be raised past the hold. It is a small rule that prevents most of the recoveries above from ever being needed.