TL;DR
- Aim for under 2% hard bounces per campaign. Sending platforms and most deliverability teams agree on that.
- Over 5% is where platforms start pulling the brakes, and where reputation damage builds quickly.
- New domains are the most fragile. Verify every list before it touches a fresh domain.
What the sending platforms enforce
| Platform | Built-in bounce brake | Default |
|---|---|---|
| Instantly | Pauses the campaign above 5% bounces, after 200 sends | On |
| Smartlead | Pauses at a threshold you choose (bounces ÷ total leads) | Off |
| lemlist | Alerts at 10%, pauses at 50% after 100 leads | On |
From each platform's help center, checked 2026-10-17. See our guides to Instantly's auto-pause and Smartlead's bounce protection.
What Google and Microsoft care about
Gmail's bulk-sender rules publish a spam-complaint limit (keep it under 0.3%) but no official bounce number. Mailbox providers still treat repeated sends to addresses that don't exist as a sign of a purchased or scraped list, and it feeds directly into how your domain and IP are rated. Bounces and complaints tend to arrive together: a list that bounces a lot is usually a list that complains a lot.
The practical safe zone
- Under 2%: healthy. Smartlead reports a median of about 1.5% across its senders.
- 2-5%: warning. Fix the list before the next send.
- Over 5%: stop. Your platform probably already has. Verify everything before resuming.
Small samples mislead. One or two bounces on a 30-email day can look like a big percentage. Judge the rate over at least a few hundred sends, and remember a single bad import can push a fresh domain over the edge in a day.
How to stay under 2%
- Verify every list before import, including "verified" exports from lead databases. Why Apollo exports still bounce.
- Treat catch-alls separately. Should you send to them?
- Re-verify lists older than about 30 days, and long sequences before their later steps.
- Spread volume across warmed inboxes and domains so one bad batch can't sink everything.