Do soft bounces count toward my bounce rate?

It depends on who is counting. Amazon SES counts only hard bounces toward the rate it enforces; temporary failures like a full mailbox don't count. Your own reporting may mix both, so check how your platform defines the metric before comparing numbers.

Is a 5% bounce rate bad?

Yes. At 5% Amazon SES places an account under review, and it recommends staying below 2%. A 5% hard-bounce rate usually means part of the list is old, bought or badly captured.

Do Gmail and Yahoo publish a bounce rate limit?

Not as a single number. Their bulk sender rules set a spam complaint threshold of 0.3% and tell senders to remove invalid recipients and slow down when messages bounce.

What is a good email bounce rate? The thresholds providers publish

Sep 25, 2026 · 3 min read · Email validation

A good email bounce rate is under 2% of hard bounces per send, and lower is better. That number isn't folklore: Amazon SES recommends a bounce rate below 2%, places accounts under review at 5%, and may pause sending at 10%. Gmail and Yahoo don't publish a bounce number, but both tell bulk senders to remove invalid addresses and treat a rising error rate as a reason to send less.

There is no universal benchmark beyond that. Be wary of "average bounce rate by industry" tables that don't say how they were measured.

What the providers actually publish

Source	What it says	Metric
Amazon SES	Stay below 2%. Review at 5% or more. Possible sending pause at 10% or more.	Hard bounces only, to domains you haven't verified, over a "representative volume"
Amazon SES	Complaint rate below 0.1%. Review at 0.1%, possible pause at 0.5%.	Complaints reported through feedback loops
Google	Keep the Postmaster Tools spam rate below 0.10% and never reach 0.30%. Reduce volume when messages bounce or are deferred. Unsubscribe recipients after multiple bounces.	Spam rate; bounces as a behaviour, not a number
Yahoo	Spam rate below 0.3%. Monitor hard and soft bounces. Remove invalid recipients promptly. Use double or confirmed opt-in.	Spam rate; bounces as a behaviour

Two things stand out. First, the only hard bounce number on the list comes from an email-sending service, and it describes when that service steps in, not when inbox placement starts to slip. Second, the mailbox providers care about bounces as evidence of how you built the list. A sender with clean opt-in lists rarely bounces; a sender working through scraped data bounces constantly.

How to calculate it so the number means something

Bounce rate = hard bounces ÷ messages sent, per campaign or per rolling window.

Separate hard from soft. A hard bounce (5xx, such as 5.1.1 "mailbox does not exist") is permanent. A soft bounce (4xx, such as a full mailbox or greylisting) may succeed on retry. Mixing them inflates the number and hides the real problem. See hard bounce vs soft bounce for the codes.
Measure per source. One rate for the whole account averages a clean newsletter list with a bad import. Break it down by list, segment or lead source.
Watch the first send to a new list. That's where bounces cluster, because nothing has been filtered yet.

What drives bounces up

Old addresses. Work addresses die when people leave a company or a company changes its domain. A list that was clean a year ago isn't clean now. B2B email list decay covers how fast this happens and how often to re-check.

Imported and purchased lists. Data you didn't collect yourself has unknown age and unknown accuracy. SES lists "don't buy, rent, or share email addresses" as its first recommendation for reducing bounces.

Typos at capture. gmial.com, hotmal.com, a missing letter in a company domain. Every typo is a future bounce, unless the domain is registered and accepts everything, in which case it might be a spam trap instead.

Disposable addresses. Many throwaway inboxes are designed to expire soon after use, so they bounce when you mail them later.

No bounce handling. If hard-bounced addresses aren't suppressed automatically, every campaign re-sends to them.

How to get under 2% and stay there

Confirm addresses at signup with a link or code. Both Google and Yahoo recommend confirmed opt-in.
Catch typos while the user is still on the form. Verifying email at signup shows how to prompt "did you mean" without blocking anyone.
Verify imported lists before the first send, and re-verify segments you haven't mailed in months.
Suppress every hard bounce immediately, in every tool that sends.
Send to risky segments (catch-all domains, long-dormant contacts) in small batches, separately from your main sends.

Step 3 is where a verifier earns its cost. pingutil's /v1/verify/batch takes up to 50 addresses per request and returns valid, invalid, risky or unknown per address, without sending a message. Catch-all domains come back risky (accept_all), and greylisting or blocks come back unknown rather than a guess, so you can see how much of the list is uncertain before you mail it. The email verification page has a demo, and the API reference lists the fields.

If a send comes back above 2%, stop mailing that source, find out where the addresses came from, and verify the rest before you resume.

Questions people ask

Do soft bounces count toward my bounce rate?: It depends on who is counting. Amazon SES counts only hard bounces toward the rate it enforces; temporary failures like a full mailbox don't count. Your own reporting may mix both, so check how your platform defines the metric before comparing numbers.
Is a 5% bounce rate bad?: Yes. At 5% Amazon SES places an account under review, and it recommends staying below 2%. A 5% hard-bounce rate usually means part of the list is old, bought or badly captured.
Do Gmail and Yahoo publish a bounce rate limit?: Not as a single number. Their bulk sender rules set a spam complaint threshold of 0.3% and tell senders to remove invalid recipients and slow down when messages bounce.

Keep reading

B2B email list decay: why work addresses go stale and when to re-verify
Work email addresses die when people change jobs and companies change domains. How B2B email list decay happens, how to spot it, and a re-verification cadence.
Hard bounce vs soft bounce: SMTP codes and what to suppress
Hard bounces are permanent 5xx failures, soft bounces are temporary 4xx ones. How to read SMTP and enhanced status codes, and which addresses to suppress.
How to verify email at signup without adding friction
Verify email at signup in real time: catch typos with a did-you-mean prompt, block only what is certainly bad, and warn instead of blocking on everything else.