List hygiene is the single highest-leverage deliverability activity most teams ignore. The instinct is to grow the list: every removed address feels like a lost opportunity. The math runs the opposite direction. A list with high bounce rates and complaint rates damages reputation across the entire domain, causing every other recipient — even your most engaged — to receive your emails worse. Cutting the list ruthlessly is one of the fastest deliverability wins available.
Bounce rate over 2% damages reputation. Complaint rate over 0.1% damages reputation severely (Gmail's threshold). Suppress hard bounces immediately. Sunset unengaged contacts at 90/180 days. Sign up for feedback loops. Re-validate quarterly. A 30% smaller, clean list out-delivers a bloated dirty one every time.
Why list quality is reputation
Mail providers evaluate sender reputation per-domain (and per-IP for dedicated infrastructure). The signals they use:
- Bounce rate. Sending to invalid addresses signals you haven't maintained your list — characteristic of bought lists, scraped lists, or lists left to rot.
- Complaint rate. Recipients hitting "mark as spam" is the strongest negative signal available.
- Spam-trap hits. Catastrophic — sending to a known trap address can blacklist your domain on Spamhaus or Validity in minutes.
- Engagement decay. Sending to dead contacts (no opens, no clicks for 6+ months) signals broadcast pattern.
Each of these compounds. A list with all four problems triggers defensive action from providers — your good recipients start receiving your emails in spam too.
Hard vs soft bounce handling
Bounces are not all equal:
- Hard bounce. Permanent failure. Address doesn't exist (5xx SMTP code, "User unknown", "Mailbox does not exist"). Suppress immediately, forever.
- Soft bounce. Temporary failure. Mailbox full, server unavailable, message too large (4xx SMTP). Retry per ESP policy; suppress after 3-5 consecutive soft bounces over a rolling window.
- Block bounce. Recipient server actively rejected based on content or sender reputation (5xx with policy message). Different problem — investigate before retrying.
Most ESPs handle this for you, but the default thresholds vary. Audit your suppression list — addresses that have hard-bounced once should never receive another send, ever, even if a user re-enters them in your form.
Suppression list hygiene
Your suppression list is as important as your sending list. It contains:
- Permanent hard-bounce addresses.
- Unsubscribed addresses — required by CAN-SPAM, GDPR, CASL. Must persist across all future campaigns.
- Marked-as-spam addresses (from feedback loops, see below).
- Manually flagged addresses (legal request, customer support escalation).
- Sunsetted addresses — long inactive contacts removed per policy.
The suppression list should override every send action, including transactional, with rare exceptions for legally-required transactional types (password reset, etc.). Re-introducing a suppressed address through any path is a bug.
Sunset policies
A sunset policy is a rule that removes addresses that haven't engaged for X days. The standard thresholds:
- Aggressive (B2C, transactional): Remove or pause at 90 days of no opens / clicks / replies.
- Standard (B2B, marketing): Pause at 180 days of no engagement.
- Lenient (newsletter, brand-awareness): 365 days maximum — anything older is dragging your reputation.
Sunset doesn't mean delete — it means remove from active sending lists. You can attempt one re-engagement campaign before final suppression (see our piece on re-engagement). After that, suppress.
Why this works: providers heavily weight whether a recipient actually engages with your sender. A list of 100K where 70K haven't opened in two years averages your engagement rate toward zero. Cutting those 70K raises your engagement rate immediately, which improves placement for the remaining 30K.
Complaint feedback loops
Major providers offer feedback loops (FBLs) that notify you when a recipient hits "mark as spam" on one of your messages. Sign up for all of them:
- Yahoo / AOL: feedback-loop.com.
- Microsoft (Outlook.com / Hotmail / Live): SNDS / JMRP via postmaster.live.com.
- Mail.ru: postmaster.mail.ru.
- Gmail: no traditional FBL, but Postmaster Tools provides aggregate spam-rate data.
When an FBL report arrives for an address, suppress immediately. Don't debate; don't try to win them back. They've explicitly said this email is spam to them.
Gmail's 2024 sender-requirements update made 0.3% complaint rate the hard threshold for bulk senders, with 0.1% as the "keep it well below" target. Crossing 0.3% sustained will cause Gmail to bulk-route your sends to spam. Most teams discover this only when their reputation is already damaged.
Address validation cadence
Validation services (ZeroBounce, NeverBounce, Bouncer, Kickbox) check addresses against MX records, syntax, role-account patterns, known-spam-trap databases, and SMTP probing. When to run them:
- At collection time. Real-time validation on every new signup catches typos and disposable addresses before they enter your list.
- Pre-import for purchased / inherited lists. Always (though purchased lists should not be sent anyway — see our separate piece).
- Quarterly cleanup. Run the existing list against validation. Suppress invalids and high-risk addresses.
- After a long pause. If you haven't sent for 3+ months, validate before resuming — addresses go invalid all the time.
Segmentation by engagement
Not all engaged recipients are equal. Segment by engagement recency for differentiated treatment:
- Highly engaged (opened/clicked in last 30 days): safe to send to frequently.
- Moderately engaged (last 30-90 days): standard send.
- Lightly engaged (90-180 days): reduce frequency; monitor.
- Dormant (180+ days): single re-engagement attempt, then sunset.
This segmentation also lets you protect reputation. Send your most experimental campaigns to highly-engaged segments first — they're forgiving and won't spike complaint rates.
Role accounts and aliases
Addresses like info@, sales@,admin@, support@ are role accounts. They're typically read by multiple people, none of whom explicitly opted in to your list. Consequences:
- Higher complaint rate (more people who didn't opt in).
- Higher spam-trap likelihood (some role addresses are repurposed as traps).
- Lower engagement (no single person owns the inbox).
Standard practice: suppress role accounts from marketing sends. Keep them only for genuinely transactional / business correspondence.
Metrics to track weekly
- Bounce rate — target under 1%, hard limit 2%.
- Complaint rate — target under 0.05%, hard limit 0.1%.
- Open rate trend — should be stable or improving.
- Reply rate (cold outreach) or click rate (marketing).
- Unsubscribe rate — high rate signals targeting issue.
- List growth net of suppression — slow growth on a clean list beats fast growth on a dirty one.
Practical hygiene playbook
- Real-time validation on signup forms.
- Hard bounces → permanent suppression on first occurrence.
- Sign up for every available feedback loop.
- Quarterly bulk validation of existing list.
- Sunset policy: 90 / 180 / 365 days based on segment.
- Suppress role accounts from marketing.
- Monitor metrics weekly; act on bounce/complaint spikes.
- Audit suppression list integrity — ensure addresses can't re-enter.