Outbound Atlas

Atlas/Features/Deliverability

Bounce protection

Auto-pausing sending when bounces cross a threshold. Every leader has a campaign-level version with different maths; mailbox-level, rolling-window protection is rare.

Featuretable-stakeshigh confidence5 minupdated 2026-10-059 sources
Status
table-stakes
Who leads
Instantly, PlusVibe, Woodpecker
The read
Table-stakes as a campaign switch, but implemented with inconsistent and sometimes misleading maths; a newcomer can win on correct, mailbox-level, pre-send protection.

Bounce protection stops sending automatically when the share of hard bounces crosses a threshold, before a bad list damages the sending domains. Buyers care because one unverified list can burn dozens of mailboxes in an afternoon, and because Google and Microsoft treat high bounce rates as a sign of a spammer (Reputation and blocklists).

How it works

The sequencer counts bounces (SMTP 5xx rejections and non-delivery reports) against sends. Once enough emails have gone out, it checks the rate against a threshold and pauses, usually the campaign. The design choices are threshold, minimum sample, denominator, scope (campaign, mailbox or domain) and what happens after a resume.

ToolDefault thresholdMinimum sampleScopeAfter resumeSource
Instantly5%, on by default200 emailsCampaignNot re-paused for 7 dayshelp
SmartleadUser-set (example 3–5%)None statedCampaignManualhelp
Woodpecker (Bounce Shield)2%Rate shown after 20 emailsCampaignManual; threshold change ignored 7 dayshelp
PlusVibeUser-set (e.g. 3%)100 emailsCampaign and account rulesRecovery rampcampaign, account
EmailBisonNot published———"smart warm-up and bounce protection" (pricing)

Who does it well

PlusVibe has the most complete design. Besides the campaign switch, its account automations watch a 3-day recipient and sender bounce rate per mailbox, cut that mailbox's campaign and warmup limits (to zero if needed), run a recovery period, and then restore settings or ramp back gradually. Instantly has the safest default: on for everyone, at 5%, after 200 sends. Its Deliverability AI Agent (Hyper Growth and up) adds pausing campaigns and replacing risky accounts. Woodpecker has the strictest default at 2%.

Where implementations differ

  • The denominator. Smartlead documents its formula as "(total bounce count + latest bounce) / total_lead_count * 100", which is bounces over all leads in the campaign, not emails sent. Upload 10,000 leads, send 300, get 30 bounces (a 10% real rate), and the formula reads about 0.3%. If the help article reflects the product, the protection fires far too late on large campaigns.
  • Sample size. Instantly waits 200 sends, PlusVibe 100, and Woodpecker shows a rate after 20. A 5% threshold after 200 sends means up to 10 hard bounces before anything happens. Spread over 20 mailboxes that is harmless; concentrated on one new mailbox it is not.
  • Scope. Almost everyone pauses the campaign. A mailbox shared across five campaigns keeps sending in the other four. Domain-level protection (pause every mailbox on a domain) is not documented by any leader.
  • Alerting. Smartlead sends a webhook with campaign name, threshold, bounce count and timestamp, which is useful for agencies running alerts in Slack.

Risks and abuse

Bounce protection is a safety feature, but thresholds also reveal what vendors tolerate. Smartlead's setup guide calls under 5% the "industry standard", while Woodpecker defaults to 2%. That spread shows there is no agreed safe number (Reputation and blocklists, How filtering works). A permissive threshold lets customers keep sending to dirty lists, and a vendor whose shared IPs or tracking domains carry that traffic shares the damage (Custom tracking domains, Platform liability: what the sequencer itself risks). The real fix is upstream: verify before sending (Built-in email verification, Email verification). Catch-all domains defeat most verifiers, which is why post-send protection still matters.

So what

Every leader treats bounces as a campaign problem. They are a mailbox and domain problem: reputation accrues to the sender, not the campaign object. The vendor whose model matches the mailbox provider's model protects customers better.

What this means for an entrant

  • Compute bounce rate over emails sent in a rolling window, per mailbox and per domain, not over campaign leads. It is simple, correct, and easy to demo against Smartlead's formula.
  • Pause at the narrowest correct scope: the mailbox first, then the domain, then the campaign if the list is the cause. Throttle rather than hard-stop where possible, as PlusVibe's recovery ramps do.
  • Ship it on by default with a strict threshold (2–3%) and a low first check (about 50 sends), with the reasoning shown in-product.
  • Bundle pre-send verification and catch-all risk scoring, and block lists that fail before the first send (Built-in email verification).
  • Expose bounce events by webhook and API (API, webhooks and MCP). Agencies wire these into their own monitoring.
9 sources cited on this page · 5 domains
  1. help help.instantly.ai
  2. help helpcenter.smartlead.ai
  3. help woodpecker.co
  4. campaign help.plusvibe.ai
  5. account help.plusvibe.ai
  6. pricing emailbison.com
  7. Deliverability AI Agent help.instantly.ai
  8. (total bounce count + latest bounce) / totalleadcount 100 helpcenter.smartlead.ai
  9. under 5% the industry standard helpcenter.smartlead.ai