SES bounce rate climbingFind the source, stop re-mailing it, and watch the trend before you send again.

Not every failed delivery counts the same way. A hard bounce is a permanent failure: the address does not exist, the domain is gone, or the mailbox provider has rejected it outright. A soft bounce is temporary: a full mailbox, a server briefly down, a message too large. SES retries soft bounces on its own schedule; the enforcement number AWS watches is built from hard bounces, which will never succeed no matter how many times you try.

This distinction matters because the fix is different for each. A soft bounce needs no action from you; SES already handles the retry. A hard bounce needs the address removed from every list that could mail it again, permanently, because retrying it accomplishes nothing and every future attempt adds to the same rate.

The rate is computed over a rolling window of recent sending volume, not a lifetime total. That cuts two ways: a single bad campaign can spike the number fast on a low-volume account, and a genuinely fixed list will see the spike age out of the window as clean sending accumulates behind it. Waiting for that alone is slow and risky; stopping new bounces is what actually moves the number.

1. Check bounce counts per campaign, not the account total. The spike almost always traces to one send, not a slow account-wide drift.
2. Look first at anything imported recently. Old exports from a prior tool, list purchases, and CSVs nobody has validated in years are the usual sources of hard bounces.
3. Check re-engagement or win-back sends specifically. Mailing a segment that has not opened anything in a long time is a common, avoidable spike.
4. Confirm whether the bounces are hard or soft before reacting. A soft-bounce cluster from one recipient domain having an outage is not the same problem as a hard-bounce cluster from a stale list.

A hard bounce should suppress the address across every list and every send path at once, not just the campaign tool that triggered it. The failure mode that keeps bounce rates high is a suppression check that only covers campaigns while a transactional API or a manual test send can still reach the same dead address. If a re-import later brings the same address back in from a fresher export, it should stay suppressed rather than resurrect.

Open the bounce chart, or a health view showing rate against AWS's thresholds, before you schedule the next campaign, not after. If the rate is still high, send to your cleanest, most recently engaged segment first rather than the full list. A clean send narrows the rolling window's average back down; another broad send while the rate is still high does the opposite.

SendBunny captures bounce events from SES automatically and suppresses hard bounces the moment they happen, enforced on campaigns, the transactional API, and test sends alike, so a re-import cannot resurrect an address that already failed. Bounce trends sit on the dashboard next to every campaign and in a health strip visible on every screen, so the number is in view before you schedule the next send rather than after AWS opens a review. It is $199 once, installed into your own AWS account.