Home › Blog › Cold Email Bounce Stop Rules: What to Do at 1%, 2%, 3% and 5%

Back to Blog
Operations5 min readJuly 19, 2026

Cold Email Bounce Stop Rules: What to Do at 1%, 2%, 3% and 5%

Reddit says pause at 0.5%. Also at 2%. Also 8%. The confusion comes from mixing three different bounces and two different campaign types into one number. Here's the tiered rule we run at 1M+ sends a month, including where the real hard stop sits, and why risky segments get a stricter one.

Cold Email Expert · WarmInboxes
Cold email bounce rate thresholds and stop rules

Quick Answer: The Tiered Rule (Hard Bounces Only)

<2% healthy · 2-3% watch + read the codes · 3-5% investigate actively (brief fresh-list tolerance) · 5% = hard stop. Risky segments, catch-all quarantines, unproven sources, stop at 3%. Count hard bounces only; soft-bounce walls are a throttling signal with a different response. And the number never acts alone: the bounce codes tell you which emergency you actually have.

Why the Internet Can't Agree on One Number

Three honest confusions produce the 0.5%-to-8% spread. First, different bounces: some people count everything including deferrals; stop rules should count hard 5xx only. Second, different channels: sub-1% norms come from opted-in marketing email, where the list consented and decay is low, cold B2B data structurally carries more risk even when verified. Third, different campaign types: a fresh unproven source deserves a tighter leash than a long-running campaign on a proven list. One number can't serve all of that; a tiered rule can.

The Tiers, and What Each One Means Operationally

  • •Under 2%, healthy. Normal cold email on verified data. No action beyond the standing hygiene: suppress every bounce permanently, keep the weekly seed test running.
  • •2-3%, watch. Not an emergency, but pull the codes: a drift of user-unknowns means the list is aging faster than expected (or a catch-all slice slipped through); any policy-block wording jumps you straight to the reputation track regardless of percentage.
  • •3-5%, investigate now. Something is wrong; you're diagnosing while exposure continues, so move fast. Acceptable briefly on an isolated fresh-source test; unacceptable as a trend on your main sends. Identify the segment driving it and cut that segment first, surgical beats total.
  • •5%, stop. Pause the campaign, full autopsy: codes classified, remaining list re-verified, source reviewed, domain health checked (blacklists + seed test). Resume only at half volume with a verified remainder. This is the fleet-wide incident line, past it, every send is spending reputation you'll need weeks to earn back.
Why risky segments stop at 3%: a catch-all quarantine or an untested data source starts with a known-elevated failure probability, that's why it's isolated in the first place (the catch-all policy). Isolation only protects you if the tripwire is tighter than the main fleet's: the experiment gets 3%, the proven campaign gets 5%, and neither ever shares infrastructure with the other.

The Part That Matters More Than the Threshold

Stop rules are the seatbelt; the driving is everything before the send. The operations that live under 2% do the same four things: verify at send-time regardless of source badges (why database "verified" isn't), resolve catch-alls with a second verifier and discard the unknowns, clean at import (dedupe, role-address stripping, the CSV Cleaner does it free), and suppress every bounce forever so no address ever bounces twice. Do those and the stop rules become the alarm you never hear, which is the point of an alarm.

Monitoring Cadence

Check bounce rate daily for a campaign's first week and after any new list injection; weekly on stable campaigns. Measure per-campaign AND per-domain, a fine 1.5% average can hide one domain at 6%, and the domain is what burns. First 48 hours of any fresh source is the danger window: that's when stale data announces itself, and catching it at send #200 instead of #2,000 is the whole difference between an autopsy and a recovery program.

Frequently Asked Questions

At what bounce rate should I pause a cold email campaign?

Our fleet rule, counting HARD bounces only: under 2% is healthy, 2-3% means watch closely and check the codes, 3-5% means investigate actively (tolerable briefly on a fresh list while you diagnose), and 5% is the hard stop, pause the campaign, autopsy the list, re-verify before resuming. Risky segments (catch-all quarantines, experimental sources) run a stricter 3% stop because they've earned less benefit of the doubt.

Does the bounce threshold count soft bounces?

No, stop rules are about hard (5xx permanent) bounces, which are what damage reputation when they accumulate. Soft bounces (4xx deferrals, mailbox-full) are mostly weather and retry logic; a WALL of soft bounces from one provider is a throttling signal with its own response (halve volume there), not a campaign-pause trigger. Know which you're counting before applying any rule.

Why do people recommend pausing at 0.5% or 1%?

They're usually quoting marketing-email norms, where opted-in lists genuinely should bounce near zero, or being conservative to be safe. Cold email to even well-verified B2B data carries structural bounce risk (data decay, catch-alls) that makes sub-1% an aspiration, not a stop rule. Pausing every campaign at 1% means living in permanent pause; the discipline that matters is the 3-5% investigation band and the 5% stop.

Is a 5% bounce rate on a fresh list ever acceptable?

As a transient on day one of a new source, briefly, while you're actively diagnosing, on an isolated segment. Never as a steady state, and never blended into your main sends. If a fresh list opens at 4-5%, the verification step failed: pause the source, re-verify, and read the codes to see whether it's stale data or unresolved catch-alls before another send goes out.

What do I actually do when I hit the stop?

Pause the campaign (not necessarily the whole domain), pull the bounce codes and classify them (list vs reputation vs throttling, see the bounce-codes decision tree), re-verify the remaining unsent list, permanently suppress every bounced address, and check the domain's health (blacklists + seed test) before resuming at half volume. Resuming without the autopsy just schedules the next stop.

Get FREE .com Domains

Prewarmed Inboxes So You Can Send Cold Emails Today

Get Prewarmed & Fresh Inboxes

Prewarmed & Fresh Inboxes

Google, Microsoft 365 and Azure inboxes with free prewarmed domains and DNS configured before you log in. Prewarmed inboxes send the day they land; every order is delivered within 48 to 72 hours.

Get Prewarmed & Fresh Inboxes