How often should you clean your email list? (and what "clean" actually means)

The short answer: clean every 3–6 months for most senders, monthly for fast-growing or high-volume lists, and always before a major campaign or after a long sending pause. If your bounce rate is climbing, you don't have a sending problem — you have a list-decay problem, and no amount of template tweaking fixes it.

I learned this the expensive way. After a weeks-long pause in sending, my first campaign back landed like a brick: bounce rate spiked, the account got flagged, and months of list growth evaporated in an afternoon. The list had gone stale while I wasn't looking. What follows is the schedule I wish someone had handed me first — and what a clean actually covers, in plain English.

The consensus cadence (third-party guides, not my opinion)

The deliverability guides roughly agree, so here is the synthesis with sources:

• Every 3–6 months: a bulk clean for most senders. A 2026 deliverability checklist (updated about 5 days before I read it) puts "bulk list clean every 3-6 months" as the baseline, with monthly for lists that grow quickly through public signup forms.
• Every 6 months: an email-hygiene guide (last updated ~85 days ago) says at least every six months for active lists — and that any list dormant six months or longer should be validated before its next send, regardless of schedule.
• Quarterly for B2B: fast-growing and high-turnover lists get monthly or quarterly treatment; high-volume senders clean monthly.
• Unengaged subscribers: review at 6–12 months, then re-engage or remove (list-hygiene best-practice guide, updated about 9 days before I read it).

The trigger list that says "clean now, not next quarter":

• Bounce rate consistently above 1.5–2% (the "above 2% triggers warning-level scrutiny" rule of thumb is from a bounce-hygiene guide updated ~87 days ago — not a platform rule, a third-party benchmark).
• Open rates declining month over month.
• You just imported a batch of new addresses.
• You haven't sent in 6+ months.

What "clean" actually means: five buckets

A proper pre-send screen sorts every address into five categories:

• deliverable — safe to send to
• risky — accept-all domains, recent bounces; send with caution
• catch-all — domain accepts everything; treat as risky, not verified
• disposable — throwaway addresses; suppress
• invalid — malformed, dead, or hard-bouncing; suppress

And one honest boundary I hold on every copy of this page: mailbox-level SMTP probing is not part of a standard pre-send screen — we don't sell stages we haven't built. The screen above is the offline pipeline: syntax, disposable, role, duplicate, and suppression checks, delivered as a flagged-address report with reasons. Your sender setup (SPF/DKIM/DMARC) is checked against public DNS records. That's the v0.1 scope, stated up front — the offline screen finds the rot you can find without sending anything; it does not and cannot prove what a mailbox will do with your mail.

What goes stale first

The addresses that decay fastest are the ones you acquired fastest: contest entries, trade-show scans, purchased-adjacent imports, role addresses (info@, support@) on B2B lists. The guides are consistent that high-turnover B2B lists decay far faster than consumer lists — job changes kill addresses monthly, not yearly. If your list is mostly B2B, quarterly is the floor, not the aspiration.

Also: hard bounces get removed immediately (not at the next quarterly clean — now), and any list that sat 6+ months unused gets a full validation before its first send. Those two rules would have saved my account.

When cleaning alone isn't enough

A clean schedule keeps a healthy list healthy. It does not rescue a list that's already in trouble — one where you suspect a large fraction of addresses are dead, and sending the next campaign feels like the thing that will get you flagged. That's a pre-send screen decision, not a hygiene schedule.

If you want someone to do that screen for you: a fixed-price ListProof pre-send audit, 7-day turnaround. You get the five-bucket breakdown, a stated prediction figure, and your sender setup checked against public DNS records. The promise is deliberately weak and bounded: if the prediction misses, we re-run the analysis free — within 30 days of receiving the findings. No guaranteed bounce rate; this is a forecast, not a crystal ball.

And if the screen says you need a full rescue, the audit fee is credited in full toward the rescue tiers within 30 days. Manual invoice; current pricing on request.