Free resource · use it on this page

Segmentation Planner

Design attribute, behavioral, and lifecycle segments on paper before you build them, so every message has a reason to reach the person who gets it.

Why plan segments before you build them

Most segmentation problems are design problems, not tooling problems. A segment built in five minutes inside a dashboard usually has an entry rule, no exit rule, and no answer to the only question that matters: what does this group get that everyone else does not? This planner makes you answer that question before you touch any software.

Every useful segment belongs to one of three families. Attribute segments are built from things a customer is: their location, plan, tier, or a preference they told you. Behavioral segments are built from things a customer does: visit frequency, spend, category purchases. Lifecycle segments are built from where a customer is in their relationship with you: new, active, at risk, lapsed. The families need different data, go stale at different speeds, and deserve different messages, so name the family before you name the segment.

Work through this page top to bottom. The example table gives you reference points, the two worksheets are the actual planning work, and the checklist is the gate before you build anything. Fill the worksheets on screen or print the page and do it with a pen. One completed segment worksheet per segment; if you cannot complete a line, that segment is not ready to exist.

Reference: well-formed segments from all three families

These are patterns, not prescriptions. Notice that every row has an exit rule and a distinct message. That is what separates a segment from a list.

Example segmentFamilyEntry rule (example)Exit rule (example)What to send
New this weekLifecycleFirst visit or pass install within the last 7 daysThe 7 days elapseA welcome and one clear next action, nothing else
RegularsBehavioral3 or more visits in a rolling 30-day windowFalls below 3 visits in the windowRecognition and early access, not discounts they would have come in without
At riskLifecycleWas active, now silent longer than your at-risk thresholdAny visit, or crossing into lapsedOne honest nudge, then quiet
LapsedLifecycleNo activity past your lapsed thresholdAny activity restarts the lifecycleA single win-back offer with a real deadline
Top tier by spendBehavioralSpend above a tier line you define, measured over a set windowDrops below the line at a scheduled reviewPerks and previews, at a lower frequency than you are tempted to use
Near one locationAttributeHome store or area recorded at signup or installLocation field changesThat store's hours, events, and openings, never another store's
Category fansBehavioral2 or more purchases in one category90 days with no purchase in that categoryNew arrivals in that category only
Plan or membership tierAttributeTier field on the customer recordTier field changesBenefit reminders and renewal notices matched to the tier

Segment definition worksheet (one per segment)

Copy this block once for every segment you intend to build. A blank line is a signal: it means the segment is decorative, not designed.

Segment name

Plain language a colleague would understand at a glance. Lapsed regulars, not Seg_RFM_23.

Family

Attribute, behavioral, or lifecycle. If the rule mixes two families, split it into two segments.

Entry rule

The exact condition with numbers in it. No purchase in 45 days is a rule; inactive is an opinion.

Exit rule

What removes someone. Every segment needs one, or it only ever grows.

Data source

The specific field or event the rule reads, and whether you actually collect it today. Not could collect. Do collect.

Estimated size today

A head count, even a rough one. Below about 50 people, treat any campaign result as an anecdote, not a signal.

Message promise

The one thing this segment receives that no other segment does. If you cannot state it, merge this segment into another.

Frequency cap

Maximum sends per month to this segment across all campaigns combined, not per campaign.

Success metric

One number you will check: redemptions, return visits, or simply an opt-out rate that stays flat.

Review date and owner

A named person and a calendar date to check size and results. Unowned segments drift until nobody trusts them.

Lifecycle thresholds worksheet

Lifecycle stages are defined by silence, and the right amount of silence depends on your visit rhythm. A coffee shop and a tire shop have completely different meanings of lapsed. Fill in your own numbers; the hints are starting points to adjust, not research findings.

New customer window

Days after first visit or install during which someone is treated as new. A week or two is a common starting point; long enough to welcome, short enough to move on.

Active definition

The action that counts as activity (visit, purchase, scan) and the rolling window. For example: any visit in the last 30 days.

At-risk threshold

Days of silence before a nudge. A workable starting point is one and a half to two times the typical gap between visits for your business.

Lapsed threshold

Days of silence before win-back mode. Usually two to three times the at-risk number. At-risk and lapsed must not overlap; a person is in exactly one stage.

Top-tier entry

The visit count or spend level that earns your best-customer tier, measured over a stated window such as 90 days or a year.

Top-tier exit

How someone leaves the tier without feeling punished. A scheduled annual review reads as fair; instant demotion after one quiet month does not.

Stop-sending cutoff

The silence length after which you stop messaging entirely. One respectful goodbye send is kinder than years of ignored noise, and it protects your deliverability everywhere else.

Pre-build checklist: pass every item before you build

Run each planned segment through this list. Any unchecked box is a reason to fix the plan, not a reason to build anyway.

  • Every segment has an exit rule

    A segment with no way out is a mailing list, and mailing lists rot.

  • Lifecycle stages do not overlap

    One person sits in exactly one stage. If someone can be both at risk and lapsed, the thresholds are wrong.

  • Entry rules use numbers, not adjectives

    Engaged and loyal are opinions. Visited twice in 30 days is a rule someone else could apply and get the same list.

  • The data behind every rule actually exists

    A rule written against a field you never populate selects nobody, silently. Check the field before you write the rule.

  • Each segment earns a different message

    If two segments would receive the same campaign this quarter, merge them until they diverge.

  • Small segments have a stated purpose

    A 30-person top tier is fine for perks. It is the wrong place to measure whether a campaign works.

  • Overlap conflicts are decided in advance

    When one person qualifies for two campaigns in the same week, which one wins? Write the priority order down now.

  • Frequency caps are set per segment

    Caps per campaign still allow pile-ups. The cap that protects the customer is the total across everything you send.

  • There is a plan for people who match nothing

    The unmatched group still exists. Decide what they get, even if the deliberate answer is nothing.

  • At least one segment is defined in order to stay quiet

    Marking a group as do-not-message is segmentation too, and often the highest-trust move you can make.

  • The channel matches the segment

    Silent updates suit everyone; interruptions suit intent. On a wallet platform like PushNotice, the pass face can update for all holders while lock-screen messages stay reserved for segments such as at-risk nudges.

  • Opt-outs are honored per person, not per segment

    Someone who opted out of win-back offers did not opt back in by becoming a regular again.

Build order: which segments to create first

Do not build the whole plan at once. Build in this order and let each layer prove itself before adding the next.

  1. 1

    Start with lifecycle: new, active, at risk, lapsed

    Four stages fall out of a single data point you already have: the date of each customer's last activity. This is the highest-value segmentation for the lowest data cost, and every later segment sits on top of it.

  2. 2

    Add one behavioral segment tied to revenue

    Usually a regulars or frequency tier built from visit or purchase counts you already log. Resist adding a second behavioral segment until the first one has received a campaign and you have read the result.

  3. 3

    Add attribute segments only when a campaign needs one

    Location, tier, and preference fields are easy to collect and easy to never use. Wait until a concrete campaign requires the split, then add exactly the attribute it needs.

  4. 4

    Run one full cycle before expanding

    Let a complete visit cycle pass and watch people move between stages. If nobody ever exits at risk, the threshold is wrong, not the customers. Fix thresholds from the worksheet above before inventing new segments.

From the guide: Best Push Notification Tools (2026)

This resource accompanies the full article — worth reading before you commit to a tool.