Segmentation Mistakes Founders Make and How to Fix Them
Segmentation mistakes founders make: nine patterns that turn a contact base from an asset into a landfill, and the small habit that repairs each one of them.
The segmentation mistakes founders make are nine patterns that turn a contact base from an asset into a landfill, and every one of them is a choice that can be reversed. None is a failure of technology. They are failures of definition and upkeep.
Definition takes an afternoon and upkeep takes thirty minutes a month, and both are available to anyone this week. This post walks through the nine, shows how to recognise each in your own base, and gives the small habit that repairs it.
Segmentation mistakes founders make: the nine at a glance
| Mistake | What it looks like | The repair |
|---|---|---|
| Sending everything to everyone | One message for all | Write for the groups you define |
| Segmenting on demographics | Cuts by age and city | Use fit and recent behaviour |
| Tagging feelings | Tags such as hot or promising | Tag facts, express judgements as rules |
| Building fields out of tags | One tag per country | Put one value per contact in a field |
| Growing the vocabulary without a convention | Five spellings of one tag | One agreed shape, with an owner |
| Using a static list as current | Old export mailed again | Use rules for anything ongoing |
| Trusting segment size | Reports numbers never sent | Look at the delivery count |
| Renaming without checking dependants | Automations match nobody | Check what depends on a tag first |
| Never retiring anything | Dormant contacts, orphan tags | Retire as deliberately as you create |
Mistake one: sending everything to everyone
One message for all is tuned for none. The disengaged majority drags down how much of your mail arrives for the engaged few, because weak engagement teaches inbox providers that your mail is unwanted.
The repair begins small. Start with a handful of groups with distinct needs, as in how to choose your first segments, and write something different for each. Sending the identical newsletter to five segments is not segmentation.
Mistake two: segmenting on demographics because they were easy to collect
Age and city describe people. Fit and recent behaviour predict what they will do. Two people of the same age in the same city can be evaluating, never going to buy, and already a customer. The repair is the grid in how to segment by fit and behaviour: sort by fit and by recent action, and the demographic fields become a footnote.
Mistake three: tagging feelings
Hot and promising are true for a week and wrong forever after. Nothing removes them, and two people apply them to different standards. Tag facts that stay true, such as attended the onboarding session or requested a quote, and express any judgement as a rule over those facts so it updates itself.
Mistake four: building fields out of tags
Country, stage and source belong in fields, where each contact has one value. As tags they multiply, drift and contradict the real field. The rule and the repair are in tags vs fields in a contact database.
Mistake five: letting the vocabulary grow without a convention
Five spellings of the same tag mean every segment built on one of them is silently missing the other four. The fix is a single shape, a category then a value, with a one line description for each tag and a named owner for each category. The test for any new tag is whether anything will be sent, filtered, scored or counted differently because of it.
Mistake six: using a static list as though it were current
An export from months ago contains people who bought, left and unsubscribed since. Rules stay current and lists do not. Use dynamic segments for anything ongoing, and keep static ones for moments, as set out in dynamic vs static segments.
Mistake seven: trusting the segment size instead of the delivery count
The gap between the number who match and the number who will receive is made of consent, suppression and duplicates. Ignoring it means reporting numbers you never sent. Look at both numbers before every send, understand the gap, and treat a large or growing gap as the base asking to be cleaned.
Mistake eight: renaming a tag without checking what depends on it
Scoring rules, sequence triggers and saved segments keep running after a rename. They report no error and they match nobody. A tidy minded person renames a tag and three automations go dark.
- Before renaming or deleting, find out what depends on the tag.
- After any change, look for tags in use that feed nothing.
- Do this on a schedule, not only when something breaks.
Mistake nine: never retiring anything
Dormant contacts, finished segments and orphan tags each cost a little. Together they make the whole base untrustworthy. The treatment for silent contacts is in how to retire dormant contacts: one win-back, then a rest. For segments, delete any that nobody can explain.
Which mistakes are linked
The nine are not independent. Several cause one another, and fixing the root removes more than one symptom.
| Root cause | Mistakes it produces |
|---|---|
| No decision about fields and tags | Fields built from tags, tags for feelings |
| No convention or owner | Vocabulary growth, broken renames |
| No upkeep rhythm | Never retiring, stale lists, growing gaps |
| No written definition of a good lead | Demographic cuts, one message for all |
This is why the guide's advice is to settle the scheme before you tag anything. An afternoon of decisions prevents a long tail of repairs.
A quick self-check
Read the nine and mark honestly how many you recognise in your own base. A count of any size is normal. The useful move is to pick the one that costs you most today, not to repair all nine at once.
- If your sends are broad, start with mistakes one and two.
- If your tags are untidy, start with three, four and five.
- If your numbers do not match what arrives, start with six and seven.
- If your automations are quiet, check eight.
- If the base feels heavy, start with nine.
For the overall picture of how the pieces fit, what is contact segmentation is the place to begin.
The monthly pass that prevents most of them
A thirty minute hygiene pass, once a month, is the difference between a base you trust and one you dread opening. It covers tags with almost no contacts, tags nobody can explain, new duplicates, stale stages, the gap between size and recipients, and segments that no longer have a purpose. None of it is difficult and all of it is easy to postpone.
A base cleaned monthly stays trustworthy. One cleaned annually becomes a project nobody wants.
What each mistake costs you
It helps to know what you are paying for each habit, so the cheapest repair can go first. None of these costs appears on an invoice, which is why they last.
| Mistake | What you pay |
|---|---|
| Sending everything to everyone | Weaker engagement, and fewer of your messages arriving |
| Demographic segments | Groups that describe people but do not predict what they will do |
| Tagging feelings | Labels that are false soon after they are applied |
| Fields built out of tags | Two records that disagree about the same person |
| No convention | Segments silently missing part of their audience |
| Static list as current | Mail to people who have bought, left or unsubscribed |
| Trusting size over delivery | Reports of numbers never sent |
| Renaming blind | Automations that run and match nobody |
| Never retiring | A base nobody trusts |
How to spot each one in an hour
You can test for most of these without a project. Set aside a single sitting and run through the quick checks below, noting each answer as you go.
- Open your last three sends. Were they the same message to the same audience?
- Read your segment definitions. Do they mention fit or recent action, or only fields such as age and city?
- Scan your tags for words that express an opinion.
- Look for tags that repeat a field, such as a tag per country.
- Look for near-duplicate tag names.
- Find the oldest static list still in use and ask who has left it since.
- For your next send, write down the matching count and the delivery count.
- List what depends on your five most used tags.
- Count the segments nobody can explain.
Each question maps to one mistake. The answers will not be flattering, but they are specific, and a specific problem is a problem with a repair.
Why founders make them anyway
None of these mistakes comes from carelessness. They come from the default. The default is to send to the whole table, to use whatever fields the form collected, to add a tag when a thought occurs, and to leave things where they are. Each is the easiest thing to do at the moment it happens.
The guide's answer is not more effort but earlier decisions. Settle the vocabulary before you need it. Define each segment in one sentence. Make the system apply the rules, so that discipline is not a matter of remembering. A scheme designed in an afternoon outlives one improvised over a year, and is far easier to trust.
Frequently asked questions
Which mistake is the most common?
Sending everything to everyone is the default, so it is the one almost every young company starts with. It is also the most expensive, and it never shows up on an invoice.
Can I fix several at once?
It is better to fix the one that costs you most, then the next. The mistakes are linked, so fixing a root cause such as the missing convention often removes more than one symptom.
Is any of this a technology problem?
No. They are failures of definition and upkeep. A spreadsheet, a written convention and real discipline are enough to avoid all nine.
How do I know the repair worked?
Look at the numbers by segment. If a group now responds differently, the count matches what is delivered, and nothing in the base surprises you, the repair has worked. The reading method is in how to measure results by segment.
The full method, with the checklists, is in The Segmentation Playbook, a free 35-page playbook.