What is a good bounce rate for cold email?

There is no official threshold. Our working rule is under 2% hard bounces is healthy, 2% to 5% needs attention before the next send, and over 5% means pausing that list.

What is the difference between a hard and a soft bounce?

A hard bounce is permanent: the address does not exist, such as Gmail's 550 5.1.1 error. A soft bounce is temporary, such as a full mailbox or a server that is briefly unavailable.

Does Gmail penalise senders for bounces?

Gmail's sender guidelines tell senders to automatically unsubscribe recipients who have multiple bounced messages, and require all senders to keep reported spam rates below 0.3%. Repeatedly sending to dead addresses is a pattern of an unmaintained list.

Why are my verified emails still bouncing?

Verification is a check at a point in time. People leave jobs, catch-all domains accept everything at verification and reject later, and some verifiers mark unknown results as valid.

Cold email bounce rate: what is too high and how to fix it

How to calculate it

Bounce rate = hard bounces ÷ emails sent, per campaign and per sending mailbox.

Count hard bounces only for this number. Soft bounces matter, but they are a different problem and mixing them in makes the figure jump around for reasons unrelated to your data.

Type	What it means	Example	What to do
Hard bounce	Permanent failure. The mailbox or domain does not exist	Gmail's `550 5.1.1`: the account you tried to reach does not exist	Remove the address for good. Never retry
Soft bounce	Temporary failure	Mailbox full, server busy, message too large	Retry later. Remove after repeated soft bounces
Block	The receiving server refused you, not the address	Policy or reputation rejection	Stop and check your authentication and reputation. This is not a list problem

Why bounces cost more than the lookup

A bounced email costs you three things.

The data. Whatever you paid for the address.
The send. A slot in a mailbox you are probably limiting to protect its reputation.
The signal. Mailbox providers see you sending to addresses that do not exist. Gmail's sender guidelines tell senders to automatically unsubscribe recipients who have multiple bounced messages, and say not to send to people who did not sign up or to buy email addresses from other companies. A sender who keeps hitting dead mailboxes looks like one working from a bought list.

The first cost is often refunded. The third is not.

A worked example

Say you send 500 cold emails from a list you bought from a database.

At a 3% bounce rate, 15 bounce. Annoying, survivable.
At a 15% bounce rate (the low end of what third-party tests report for some large databases), 75 bounce.
At 35%, the top of that reported range, 175 bounce.

Now suppose your provider refunds bounced credits, as Apollo does within 30 days for its verified emails sent through Apollo. At 15% you get 75 credits back. You do not get back the 75 signals to Gmail and Outlook that this mailbox sends to addresses that are not real, and you have spent 15% of your limited daily sending on nothing.

The arithmetic says: a cheap list with a high bounce rate is expensive. A list that costs twice as much per address and bounces at 1% is usually the better buy.

Our thresholds, and what to do at each

There is no official bounce-rate threshold from Google or Microsoft. These are our working rules for cold email, where you have no prior relationship to cushion a bad send.

Under 2%: healthy. Keep removing every hard bounce immediately and carry on.
2% to 5%: re-verify before the next send. Run the rest of the list through a verifier, drop invalid and unknown results, and treat catch-all addresses with caution.
Over 5%: stop this list today. Pause the campaign, verify everything that remains, and find out where the bad addresses came from before sending again.
Sudden spike on a list that was fine: check whether the bounces are really blocks. If many messages fail with policy or authentication errors, the problem is your domain setup, not your data.

Finding the cause

Bounces cluster. Sort your bounced addresses and look for the pattern.

One source dominates. One database, one scrape, one export. Stop using it for email.
Old data. Addresses enriched months ago. People move jobs; re-verify anything older than about three months before sending. This is our rule of thumb, not a published standard.
Guessed addresses. first.last@company patterns that were never verified. These fail constantly at companies that use a different format.
Catch-all domains. The server accepted the address at verification time and rejects later, or accepts and silently drops. See catch-all email.
One domain, many bounces. The company may have changed email provider or domain. Check before blaming the data.

Prevention, in order of impact

Verify every address immediately before it enters a sequence, not when you bought it. How to do that properly is in how to verify email addresses.
Never send to "unknown" results in cold outreach.
Remove hard bounces instantly and never re-import them.
Stop sequences on a bounce. If step one bounced, steps two to four should never run. More in follow-up sequences that stop.
Keep SPF, DKIM and DMARC set up so blocks do not masquerade as bounces. Gmail requires SPF or DKIM for all senders, and DMARC for anyone sending close to 5,000 messages a day to personal Gmail accounts. See SPF, DKIM and DMARC.

How Sluice handles bounces

Every address Sluice returns is verified independently before you see it, and only verified addresses are billed. If one still bounces, the credit is refunded exactly what it cost, once. Sequences stop on a bounce, and bounce rate is shown by source in Insights, so you can see which lead source is producing bad addresses. It is not magic: a person can leave their job the day after verification, and catch-all domains remain uncertain everywhere. Details at pricing.

Next step

Pull your last three campaigns, compute hard bounces ÷ sent for each, and sort the bounced addresses by source. If one source accounts for most of them, you have found the fix.