Email Marketing Deliverability
Email List Hygiene: Cleaning Your List for Better Delivery
A cleaning schedule with defined rules, and the arithmetic showing why removing addresses usually raises total clicks.
On this page
The short version
- Removing subscribers usually raises total clicks. The engaged remainder gets better placement, and the improvement is larger than the loss.
- Hygiene is a schedule, not a project. Quarterly is enough for most lists; monthly if you acquire quickly.
- The addresses that never engage are the ones most likely to become spam traps, which is the outcome with no recovery path.
List hygiene is the deliverability work with the best return and the worst optics. Every other improvement makes a number go up; this one makes the headline number go down, which is why it gets deferred in favour of things that are less effective and more comfortable.
The case for it is arithmetic rather than principle, and the arithmetic is worth working through, because it is what makes the conversation with whoever owns the list size possible.
Why a smaller list gets more clicks
Engagement ratios drive placement. A list where a third of addresses never open depresses every ratio a provider scores you on, so mail to the two thirds who do open is more likely to be filtered.
Remove the dead third and two things happen at once. The ratios improve immediately, because the denominator shrank. And placement for everyone else improves over the following weeks, because the ratios are what placement is scored on.
The clicks lost from the removed segment are approximately zero, because they were not clicking. The clicks gained come from better placement on the people who were.
The arithmetic, worked
A 100,000-address list where 35,000 have not engaged in a year. Illustrative figures with plausible rates, to show the shape of the trade.
Illustrative model, not measured data — the placement and click rates are plausible values chosen to show the mechanism. Run the same arithmetic on your own figures before removing anything.
The rules, in order of confidence
Some removals are unambiguous and some are judgement calls. Doing them in order means the safe ones happen automatically and the arguable ones get a decision.
What to remove, and how confident you can be
| Rule | Confidence | Cadence |
|---|---|---|
| Hard bounces | Certain | Automatic, first occurrence |
| Repeated soft bounces over a window | High | Automatic |
| Role addresses that never engage | High | Quarterly |
| Obvious typos in the domain | High | At capture, and quarterly |
| Duplicates after normalisation | High | Quarterly |
| Never opened since signup, past a threshold | Moderate | Quarterly, after a win-back |
| No engagement in the last N months | Judgement | Quarterly, after a win-back |
The never-opened-since-signup group is the highest-risk one to keep, because it contains addresses that were mistyped, insincere, or already dead when collected — and those are exactly the ones that become recycled traps.
Set the window from your own cadence
"Six months of inactivity" means something different for a daily sender than for a quarterly one. Define the threshold in sends rather than months: someone who has received twelve consecutive messages without opening any of them has told you something, whether that took six weeks or two years.
Then convert to a date-based rule for implementation, because that is what platforms segment on — but derive it from the send count rather than adopting a number from an article.
Always attempt a win-back first
Removal should be the end of a sequence, not an event. A short win-back gives people a chance to stay and recovers a meaningful share of them, and it makes the removal defensible internally — nobody was cut without being asked.
It also produces information. Whether people respond to the frequency option rather than the content one tells you something about the whole programme, not just about the segment being cleaned.
A quarterly hygiene pass
- Back up the full list before anything is removed
- Confirm hard bounce suppression has been working since the last pass
- Normalise and merge duplicates
- Run the win-back sequence on the inactive segment
- Remove or archive those who did not respond
- Record how many were removed and why
- Re-check complaint and bounce rates on the next two sends
Frequently asked questions
Should removed subscribers be deleted or archived?
Archive to a suppressed segment that receives nothing. You keep the record for consent and reporting purposes, they stop affecting your ratios, and platform pricing based on contacts stored may still count them — which is worth checking.
What if opens are unreliable because of privacy protection?
Weight clicks more heavily and treat opens as supporting evidence. A subscriber with no clicks across a long run of sends is a reasonable candidate regardless of what the open data says.
How do I justify the list shrinking to someone who counts subscribers?
Show the arithmetic on your own numbers, and report engaged subscribers as the headline figure rather than total. A count that includes people who will never open anything is measuring storage, not audience.