D-18 · HYGIENE · VALIDATION
Invalid-email validation before you send
Every invalid address mailed costs reputation plus overages. This scenario layers syntax, mailbox, and risk checks before first send.
Symptom
A 40,000-address event import mails out and 11% hard-bounces: typo domains, disposable addresses, role-address spam traps, and long-dead corporate domains. The new-sender warmup absorbs the bounce wave as its first impression, placement drops for months, and overage billing charges for eleven thousand messages that could never convert. The list vendor’s “verified” claim proves to mean “formatted like an email.”
Signup flows leak daily too: plus-address abuse, bot registrations, and disposable domains accumulate a few percent monthly until the next bulk send detonates them together.
Cause
Validation has layers and teams apply only the cheapest. Syntax checks catch malformed strings but nothing else. DNS/MX checks catch dead domains but not dead mailboxes. Mailbox verification (SMTP probing or vendor APIs) catches unknown users but must be used carefully to avoid probing abuse. Risk signals — disposable domains, role addresses, known trap patterns, velocity anomalies — need policy decisions, not just scores. Skipping layers to “save validation spend” transfers the cost to reputation and overages at multiples.
Stale data decays independently: even a clean list goes bad at percent-per-month rates through job changes and domain churn, so one-time validation without ongoing hygiene merely delays the detonation.
Fix
Gate imports through full validation: syntax → DNS/MX → mailbox verification → risk classification, quarantining (not deleting) failures for review. Block disposable domains at capture for bulk newsletters, flag role addresses for transactional-only treatment, and typo-suggest common domains to rescue keystroke errors. Never mail first-time imports at full volume — validate, then warm progressively with engaged-first ordering.
At capture, add confirmation for high-risk sources (see double opt-in): validation predicts deliverability, confirmation proves intent. Both are needed for cold or incentivized traffic.
Prevention
Re-validate quarterly plus pre-send for cohorts older than six months, automate bounce-driven suppression so one hard bounce permanently retires an address, and track validation economics openly: cost per validated address against overage and reputation cost per bounced send. Validation always wins. Review source-level bounce rates monthly and cut acquisition sources whose invalid share exceeds tolerance regardless of volume promises.
Worked example
An event sponsor shares 40,000 “verified” leads. Full validation tells another story: 6% bad syntax and typo domains, 3% dead MX, 2% known disposable, plus a cluster of role addresses and spam-trap patterns from recycled lists. The team quarantines 4,400 addresses, typo-suggests 900 recoverable ones at capture, and mails 35,600 — with double opt-in for the coldest third.
Hard bounces land at 0.8% instead of 11%, the warmup stays on schedule, and validation spend of a few hundred dollars replaces thousands in overages plus months of reputation repair. Source analysis shows one sub-list contributed half the invalids; that supplier is never used again. The quarantine file, reviewed manually, even surfaces 200 VIP addresses with typos worth personal outreach.