Skip to content

Marketing & Growth Systems

Audience segmentation: a segment is a saved question

An exported list is right on the day you export it. A segment is the question written down: which customers are due a furnace tune-up, aren't on a plan and agreed to email? Asked again every day, it gives a different answer as customers book, join plans and pass dates, and it can tell you why each person is in.
Tune-up duefurnace on recordlast service over 11 monthsnot on a planemail consenteveryone elsein the segment212214213215214Mon212Tue214+3 crossed 11 months−1 booked a tune-upWed213−1 joined a planThu215+2 crossed 11 monthsFri214−1 unsubscribed

Four ways to group customers

Who they areHomeowners in Riverside with a gas furnaceThey move or replace the equipment
What they didOpened the furnace guide twice this monthA few weeks pass; behaviour fades fast
Where they areDue a tune-up, not on a planThey book, join or lapse, which is why it's re-run daily
How well they fitWater heater over 10 years old, no quote in a yearThe record is incomplete

Three segments, written down

A segment someone can read, own and question. These three feed messages and a report.

Tune-up due

Rule
furnace on record · last service over 11 months ago · not on a plan · email consent
Refreshed
daily, 6:00 am
Used by
tune-up reminder; fall campaign
Owner
marketing lead
Members today
214

How segments mislead

  1. 01

    The frozen export

    Right on the day it was pulled

    A spreadsheet copy stops changing while your customers don't.

    lists passed around instead of rules.

  2. 02

    Overlapping segments

    One person, three campaigns

    Without a priority rule, a customer lands in every group that matches.

    one customer getting three emails in a week.

  3. 03

    Too small to tell

    Eleven people

    A segment that small can't show whether a message worked.

    conclusions drawn from a handful of replies.

  4. 04

    Missing fields

    Unknown isn't no

    A customer with no equipment on record isn't a customer without a furnace. Keep “unknown” as its own answer.

    blank fields quietly leaving people out.

  5. 05

    Sensitive traits

    Not used

    Segments use service history, equipment and stated preferences, never guesses about age, health, income or family.

    fields that stand in for those.

What segmentation hands off

In

Customer records
services, equipment, plans
Behaviour
visits, opens, replies
Segment rules
written and owned
Audience segmentation
re-run → who's in, who left, why

Out

Members today
with the reason each is in
Changes
who entered or left, when
Counts
for reports and tests

Worth building when one list stops fitting everyone

Good fit

  • You send different messages to different customers, or want to.
  • Customers' situations change week to week.
  • Lists are exported and edited by hand today.

Not the right fit

  • You have a few dozen customers and know them all.
  • Service history isn't recorded yet.

Questions about audience segmentation

Isn't this just a filter on our customer list?

It is a filter: saved, named, owned and re-run on a schedule, with a record of who came and went. That record is what makes it trustworthy.

How many segments should we have?

As many as you have different things to say. Most service businesses need a handful, each tied to a message or a report.

Can AI find segments for us?

It can suggest groupings worth a look. A segment that drives messages should still be a rule a person can read.

What about customers in two segments?

That's normal. A priority rule decides which message wins, so nobody gets everything at once.