The bounce rate that decides where your email lands

Most senders find out about bounce rate thresholds after they have crossed one. Open rates fall by half in a week, replies stop, and nothing in the campaign itself has changed. The mail is going to spam, and the reason traces back several sends.

The number

Google publishes a spam complaint threshold and treats bounce rate as part of the same reputation picture. Microsoft is less specific in public but behaves similarly. In practice, sustained hard bounce rates above roughly 2 percent are where senders start seeing placement change, and above 5 percent the change is fast and obvious.

Below about 1 percent, nobody is looking at you. That is the range to stay in.

One send at 4 percent will not destroy a domain that has been clean for a year. Four sends in a row at 4 percent will.

Hard and soft bounces are not the same

A hard bounce is a permanent rejection. The mailbox does not exist, the domain does not resolve, or the server has refused the address outright. These are what damage reputation, because they signal that you are mailing a list you did not verify.

A soft bounce is temporary. The mailbox is full, the server is busy, the message was too large. These carry far less weight, though a soft bounce that repeats on the same address for a week is treated as a hard one eventually.

When a tool quotes you a bounce rate, check which one it is counting. A 3 percent combined figure with most of it soft is a different situation from 3 percent hard.

What actually produces hard bounces

In order of how often it comes up:

1. List age. B2B addresses decay at roughly 2 to 3 percent a month as people change jobs. A list that was clean in January is meaningfully dirty by June, and one from last year is mostly archaeology.
2. Scraped or purchased data. Whatever the vendor claims, assume a double-digit invalid rate until you have measured it yourself.
3. Pattern guessing. Tools that construct first.last@company.com without confirming it produce addresses that look perfect and bounce hard.
4. Typos at capture. gmai.com and hotmial.com are on every list that has a signup form and no correction step.
5. Role addresses that were retired. info@ and sales@ get decommissioned quietly, and nobody updates the CRM.

Why recovery is slow

Reputation systems are weighted towards recent behaviour but they do not reset. Once Gmail has decided your domain is a risk, you are demonstrating good behaviour against a running average that includes the bad sends.

Realistically, recovering from a serious reputation hit takes four to eight weeks of clean, low-volume, well-engaged sending. During that time your mail is landing in spam, so engagement is poor, which slows the recovery further. It is a bad loop and the only reliable exit is time.

This asymmetry is the whole argument for verification. Cleaning a list costs a few dollars and takes minutes. Recovering a domain costs six weeks of pipeline.

What to do before a send

Verify the list, not a sample of it. Sampling tells you the rate. It does not remove the bad addresses, and the bad addresses are what cause the damage.

Verify close to the send. A list verified three months ago has already drifted by 6 to 9 percent. Verify within a week of sending.

Separate the uncertain records. Catch-all domains cannot be confirmed by anyone. Keep them in their own group and send to them from a domain you can afford to damage, or not at all.

Fix typos rather than dropping them. jane@gmai.com is a real person who mistyped. Correcting the domain recovers a contact that would otherwise have bounced.

Watch the rate per send, not per quarter. A quarterly average hides the one campaign that did the damage.

What to do after a bad send

Stop sending from that domain immediately. Do not try to make up the volume elsewhere in the same week.

Verify the entire list, including the parts you were confident about. Then restart at low volume, to your most engaged contacts only, and hold it there for two weeks before increasing. Watch the bounce rate on every send and stop again if it rises.

If the domain was badly hit, moving new outbound to a separate sending domain while the original recovers is reasonable. Warm the new one properly rather than treating it as a fresh start at full volume, or you will be in the same position with two damaged domains instead of one.

The short version

Stay under 1 percent hard bounces per send. Verify within a week of sending, keep catch-all addresses in their own group, and correct typos instead of discarding them. The cost of prevention is a few dollars per thousand addresses. The cost of the alternative is measured in weeks.