What Is Contact Segmentation? A Plain Answer for Founders
What is contact segmentation? It is deciding which differences between your contacts should change what you say, when, and whether you say anything at all.
Contact segmentation is deciding which differences between your contacts should change what you say to them, when you say it, and whether you say anything at all, and then recording those differences so they stay accurate as your list grows. It is not slicing a list by age and city because the form happened to ask for them.
Every young company reaches a point where its list of contacts stops being a list of people the founder remembers and becomes a list of rows. This post explains what segmentation is, what it is not, and how to tell whether you need it yet.
What is contact segmentation, in practice?
Picture the table of contacts in most small companies. Somebody signed up from a webinar. Somebody downloaded a guide. Somebody was a customer, then left. Somebody is a competitor who wanted to see the pricing page. Somebody is a test address the developer created on a Tuesday. They all sit in the same table, and the default thing to do with that table is to send all of it the same email.
Segmentation replaces that default. You write down a small number of groups, define each group by a rule or a list, and send each group something suited to where it is. The word that matters in that definition is decide. A segment is a written decision about who belongs together and why, not a filter someone saved and forgot.
It is also a way of recording what you know. A segmented base can say which contacts fit your target, which have done something recently, and which are customers. An unsegmented base can only say that a name exists.
The habit it replaces
Sending one message to everyone is the most expensive habit in early marketing, and it never shows up on an invoice. A message written for everyone is tuned for nobody. It lands softly with the people it should have moved and irritates the people it should have left alone.
The damage then compounds in three directions.
- Engagement drops. Weak engagement teaches the inbox providers that your mail is unwanted, which lowers how much of it arrives.
- Sales loses its bearings. Nobody can tell which contacts are worth a call, so people call whoever is newest.
- Questions go unanswered. When someone asks which kind of customer converts best, nobody can say, because the base was organised to be sent to, not to answer questions.
A list you send everything to is not an audience. It is a liability you have not been billed for yet.
What segmentation is not
The word gets stretched, so it helps to mark the edges.
| It is | It is not |
|---|---|
| A decision about which differences should change what a contact receives | Dividing the list by whatever fields the form collected |
| A group defined by a rule or a deliberate list | A saved filter nobody can explain |
| Different messages for different groups | The same newsletter sent five times to five lists |
| A way to leave some people alone | A way to reach more people |
| As few groups as will change a decision | As many groups as the software allows |
The last row deserves emphasis. Segmentation multiplies work, because every segment is a message to write differently, a definition to maintain and a number to interpret. A company with a few hundred contacts and one offer needs three or four segments, not a dozen.
Six jobs a segment does
Segmentation earns its place through six mechanisms. None of them is a slogan about personalisation. Each is a lever you can pull on purpose.
- Relevance. People read a message because the first line looked like it was about their situation. A message written for one recognisable situation can be specific in a way a message for everyone never can.
- Prediction. Fit and recent behaviour tell you what someone is likely to do next. Easy fields such as location and age only describe. The detail is in how to segment by fit and behaviour.
- Protection. Not sending is a form of segmentation. Leaving out the silent, the unsubscribed and the bounced keeps your sending reputation healthy for the people who are waiting for your mail.
- Analysis. Once contacts are sorted, the base can answer questions about which source sends people who engage and which sends people who vanish.
- A shared definition. A segment is a written definition of a good lead, which both sales and marketing can test and revise.
- Timing. Behaviour has a shelf life. Segments that move as behaviour changes let you reach someone while the moment that made them interested still exists.
Why demographics are the wrong first cut
The fields that are easiest to collect, such as location, age and a free-text job title, tell you least about whether someone needs what you sell right now. Two people of the same age in the same city can be evaluating, never going to buy, and already a customer.
What predicts the next action is two things. Fit asks whether this is the kind of person or company your product genuinely serves: the right role, the right type of organisation, the right problem. Behaviour asks what they have done lately: read, clicked, visited the pricing page, filled in a form. Demographics tell you what someone looks like. Fit and behaviour tell you what they are about to do.
The prerequisite nobody mentions: a base that is clean
Here is the honest limit. A segment of a dirty base is a precise wrong answer. Duplicated, stale or mislabelled records are not fixed by segmentation. They are hidden behind a confident looking count.
So the work has two halves that depend on each other.
- Decide the scheme: which facts are fields, which are tags, what the stages are. This takes an afternoon. See tags vs fields in a contact database.
- Keep the base under it honest: duplicates, consent, test contacts and stale records. See how to prevent duplicate contacts.
Segment as little as will change a decision, and keep the base clean enough that the segments mean something.
Do you need it yet?
You are probably ready when two or more of the following are true.
- You send the same email to everyone, and you suspect some of them should not get it.
- Sales calls whoever is newest because nothing says who is worth a call.
- You cannot say which kind of customer converts best.
- Your list holds people who bought, left, signed up twice or never agreed to marketing, and you are not sure how many.
- Marketing and sales argue about what counts as a good lead.
If none of these ring true, you may simply have a small list that you still know by heart. That will not last, and the afternoon of decisions is cheaper before the list grows than after.
What the first version looks like
Start small. Most early businesses can run almost everything from a handful of groups: new and not yet contacted, engaged, cooling, dormant and customers. How to choose your first segments walks through them. Each group gets a one sentence definition, a name that describes the rule rather than the campaign, and a distinct message.
Notice that the first version contains a group you mostly send nothing to. That is intentional. A segment that excludes the disengaged is not a smaller audience. It is a healthier sender, and a healthier sender reaches the engaged people more reliably.
How you will know it works
A segment earns its keep when it does at least one of three things: it changes what someone receives, it starts something automatically, or it makes a number readable. If it does none of the three, it is a saved filter with a name.
The third is the one founders overlook. An overall open rate or conversion rate averages groups that behave completely differently. Split it by segment and two things usually appear: one group that responds far better than the rest, and one that responds terribly. How to measure results by segment covers how to read those two.
A first-week plan
| Step | What you do | Output |
|---|---|---|
| One | Write down what good fit means for your business, in facts you can record | A short fit definition |
| Two | List which behaviours signal real intent and which are noise | A short behaviour list |
| Three | Decide which facts are fields and which are tags | A one page scheme |
| Four | Define the stages by event | A stage list |
| Five | Write each starting segment as one sentence | Five definitions |
| Six | Compare the number who match with the number who will receive | A first honest count |
None of this needs software. Everything in segmentation can be done with a spreadsheet, a written convention and real discipline. Software only moves the discipline from remembering to rules.
Common first mistakes
The patterns that sink early segmentation are failures of definition and upkeep, not of technology. Sending everything to everyone, segmenting on demographics, tagging feelings and trusting segment size instead of delivery count all appear in segmentation mistakes founders make. Read that list before you build, not after.
Frequently asked questions
Is segmentation the same as personalisation?
No. Personalisation changes a detail inside a message. Segmentation decides which message a group receives, and whether it receives one. Segmentation is what lets you write for a person while sending to a group.
How many segments do I need?
Fewer than you think. A company with a few hundred contacts and one offer needs three or four. Add a segment only when an existing one is plainly too broad to write a single message for.
Do I need special software?
No. A spreadsheet, a written convention and real discipline are enough to do everything in the method. Software becomes useful when remembering the rules is the thing that fails.
What if my list is messy already?
Then clean as you go. Settle the vocabulary, choose the matching key for duplicates and mark test and internal contacts. A segment of a dirty base gives a confident wrong answer, so the cleaning is part of the work, not a separate project.
The full method, with the checklists, is in The Segmentation Playbook, a free 35-page playbook.