Email segmentation that actually earns its keep
The segments that reliably pay — recency, lifecycle, behaviour — the ones that waste your time, and why two segments beat twenty.
Performance · 2 min read · updated
Why segments beat blasts
A message aimed at everyone is calibrated for no one — it has to be vague enough to be safe for the whole list, and vagueness is what readers skim past. Segmentation is not a growth hack; it is just the observation that relevance is the strongest lever in email, and you cannot be relevant to people you refuse to distinguish.
It also protects deliverability. Sending less mail to the people least likely to want it lowers complaints and raises average engagement — both of which providers score.
Segments that pay for themselves
- Engagement recency. Opened or clicked in the last 90 days, versus not. The single highest-value split in email: it feeds your sunset policy, protects your sending reputation, and costs nothing to compute.
- Lifecycle stage. Trial versus paying versus churned; first purchase versus repeat. The same message cannot serve someone deciding whether to start and someone deciding whether to leave.
- Behaviour. Bought this category, used that feature, abandoned this checkout. What people did predicts what they will do; what they are predicts almost nothing.
- Acquisition source. The giveaway-signup cohort and the paying-customer cohort have different expectations and wildly different complaint rates. Watching them separately shows you which sources grow a list and which merely inflate it.
Segments that waste your time
- Demographic astrology. Age and gender splits on a product list usually encode a marketer's stereotypes rather than a behavioural difference. Test it if you must; expect nothing.
- Over-slicing. Twenty segments on a ten-thousand-person list means twenty audiences too small to measure and twenty variants to write. Every segment you add must earn a genuinely different message, or it is overhead.
- Segments nobody acts on. If two segments always receive the same email anyway, you have a report, not a segmentation.
Start with two
Engaged-last-90-days and everyone else. Send your normal programme to the first group, a gentler cadence and eventually a win-back to the second. This one split captures most of segmentation's deliverability value in an afternoon of setup.
Add further segments one at a time, each justified by a message you could not send without it. The goal is not a beautiful taxonomy; it is fewer irrelevant emails.
Common questions
How many segments should I have?
As many as have a genuinely different message attached, and no more. For most senders that is somewhere between two and six. The count is an output of your messaging, not a target.
Is my list too small to segment?
The engagement-recency split pays at any size, because its value is deliverability rather than personalisation. Fine-grained behavioural segments need enough volume per segment to be worth separate copy.
Check your addresses now
The free InkPigeon checker runs syntax, DNS, disposable, typo and role-account analysis on any address — instantly, no signup.