What Chesterton's fence says
Every marketing system collects things over time. A field is added to a form for a campaign; a parameter is added to links for a report; a weekly export is set up for a manager who has since moved on; a rule in the automation platform suppresses a group of contacts for a reason recorded nowhere. Years later, someone tidying up meets the thing, sees no use for it and wants it gone.
The rule says that "I can't see what this is for" is not yet a reason to remove it. Its absence of an obvious purpose is a fact about the person looking, not about the thing. The right next step is to find out why it exists. If the reason turns out to be gone, removing it is safe and sensible. If the reason is still live, the thing either stays or its job moves somewhere better before it goes.
The site's guides apply the rule without naming it. The guide to reviewing recurring marketing work writes down what each routine is for and asks the people who receive it before stopping anything. The guide to retiring a marketing tool lists every dependency, from tags to the people who still log in, before the tool is switched off. The CRM field audit finds what writes and what reads each field, and the URL parameter audit asks what each parameter does to the page and why it was added. Each guide owns its own procedure; this page describes the rule they share.
Finding the reason
The reason for a fence is rarely written down, but it nearly always leaves traces: an owner, a reader, a dependency or a record. The table sorts the common marketing fences by where those traces are.
| What is being removed | Who to ask | What to check | How to switch off before deleting |
|---|---|---|---|
| A form or CRMThe system that stores an organisation's contacts, companies, deals and interactions, and the processes built around it. field | Whoever added it; the owners of reports and workflows | Every report, list, workflow, integration and scoring rule that reads it | Hide it from forms, stop writing to it, archive it, then delete |
| A recurring report or export | The people who receive it | Whether anyone opens it; what decision it feeds | Pause it and tell the recipients; restart on request |
| A URL parameterA key and value appended to a URL after a question mark, such as a filter, sort order, session or tracking tag, that may or may not change the page. | The developer and the analytics owner | What the code does with it, which internal links carry it, what reports read it | Stop adding it to new links; leave old links working |
| A tag, pixel or script | The marketing and analytics owners | Which tools and audiences depend on its data | Pause it in the tag manager and watch the receiving tools |
| An automation rule or suppression | The platform owner; compliance where consent is involved | Which contacts it affects and why they were excluded | Disable without deleting; keep the list it used |
| A tool | Every current user | Integrations, scheduled exports, forms and stored records | Switch off integrations first; export before cancelling |
| A product feature | Product, support and the customers who use it | Who relies on it and for what job | Phase it out with notice |
Start with the cheapest trace. A question to the person who added the thing, or to the people who receive its output, often settles the matter in minutes. Where nobody answers, check the system for readers: a field with no report, list or workflow reading it is far more likely to be safe than one with an unexplained reader. A consistent Naming ConventionAn agreed pattern for naming campaigns, ads, events and tracking parameters so that reports can group them reliably. makes this search easier next time, because a field or parameter named for its purpose carries part of its own reason.
Switch off before deleting. Most of the guides above retire things in stages, so a mistake shows up as a complaint or a broken report while the change can still be reversed. The guide to retiring a feature does the same with customers: it finds who relies on the feature and for which job before setting a timeline. Before a switch-off, note a BaselineAn estimate of what a metric would have done without a given change, campaign or promotion, against which the effect is measured.: the report totals, list sizes or form completions the thing currently affects. A change in those after the switch-off is the clearest sign that the fence had a job.
Record what you found. The marketing decision log template has room for the evidence behind a change and how easily it can be undone. A line saying why a thing existed and why it was removed saves the next person from repeating the search, or from rebuilding something that was removed for good reason.
When the reason is gone, and when no reason can be found
Finding the reason leads to one of three outcomes.
The reason is gone. The campaign ended, the manager left, the tool the parameter fed was retired. Remove the thing, in stages if anything still reads it, and record why.
The reason is live but the thing is the wrong way to serve it. A free-text field that a workflow depends on, a report that one person needs once a quarter. Move the job to a better home first, then remove the old one. Removing the fence without moving the job simply breaks it.
No reason can be found. The owner has left, nothing in the system reads the thing and nobody claims it. The rule gives no answer here, and pretending it does leads to keeping everything. The practical course is a reversible switch-off, long enough to catch uses that only appear at month end, quarter end or in a particular season, followed by removal if nothing breaks.
How careful to be depends on how hard the removal is to undo. Deleting a field's data, cancelling a contract or removing a suppression that protected people who had opted out are hard to reverse. Hiding a field or pausing a report is easy to reverse. The amount of checking should follow the difference.
Worked example (illustrative)
A veterinary group clears its booking form
Corbuxley Veterinary Group, an invented group of 7 practices, wants to shorten its online booking form from 14 fields to 9. Two fields look unused: "last vaccination date" and "referral code".
The marketing lead asks the practice managers and searches the CRM for anything that reads either field. Nobody remembers adding "last vaccination date", but the search finds that the vaccination reminder emails, sent to about 2,300 pet owners a year, use it to set their send date. Removing it would have stopped the reminders quietly.
The team moves the job instead: reminders now read the date recorded at the appointment, which the practices already keep and which is more accurate. After 4 weeks of checking that reminders still go out, the form field is removed.
"Referral code" has no reader at all. The team hides it from the form for a month and tells the 7 practice managers. Nobody asks for it, no report changes and the field is archived. Both decisions are written in the group's decision log.
Limits and criticisms
Finding the reason can cost more than the thing is worth. A forgotten field in a little-used form may not justify interviews and a system search; a short, reversible switch-off is often cheaper than the investigation.
Some fences have no recoverable reason. The person who added the thing has gone, nothing was recorded and nothing reads it. The rule then gives no answer, and the decision falls back on a judgement about what a removal would risk.
It can become a reason never to remove anything. "We don't know why it's there" is a fair reason to look, and a poor reason to keep something for ever. A team that cites the rule against every clean-up ends up with the clutter the clean-up was meant to fix.
And it is a caution about removals, not a method for deciding what to keep. Knowing why something was added says nothing about whether it is worth its cost today. That needs evidence of value: who uses it, for what decision and how often.
Related frameworks
One-way and two-way doors sets how much checking a removal deserves: a removal that can be reversed in minutes needs a question and a pause, while one that deletes data or ends a contract needs the full search first. The five whys traces a failure back to its cause, which is useful after a removal breaks something, and it can also be pointed forwards, asking why a thing was added until the answer reaches the decision behind it. The marketing frameworks library lists the other planning frameworks by the decision they serve.
Common mistakes
- Treating "I can't see a use" as "it has no use". The reader may be a workflow, an integration or a quarterly report.
- Deleting in one step. Hide, pause or stop writing first, so a mistake can be undone.
- Asking only the team that wants it gone. The people who receive the output know what it is for.
- Checking for too short a time. Monthly, quarterly and seasonal uses only show up when their date comes round.
- Removing the thing without moving its job. A live reason needs a new home before the old one goes.
- Keeping everything because nobody knows. Silence after a careful check and a reversible switch-off is enough.
- Not recording the reason. The next person will repeat the search, or rebuild what was removed.
