Naming Things Well

A name is a claim about what sits behind it, and people act on it without checking. The expensive failures are the accurate-once ones: a page title written for an older version of the page, an environment called staging that stopped being staging a year ago and is still quoted in the project document.

What does a name have to do?

It has to be enough for somebody who was not in the room to act correctly without asking. Elegance is downstream of that.

Which puts the real failure somewhere unexpected. An ugly name gets questioned by the first person who reads it. The one that was accurate a year ago never gets questioned, and it keeps being quoted by people with no way of knowing it moved.

Five surfaces where a name is doing load-bearing work:

Surface	What the name has to settle	How the drift shows up
Page title in search	The answer the page gives	Positions look healthy, clicks do not arrive. One account averaged 0.11% for a month
Environment or deploy target	Which copy of the site this is	Two people quote the same URL and mean different applications
Content model field	What the record holds	Migration surfaces references with no destination in the new model
Plan or price name	What the buyer owes	The call gets spent working out the number instead of the decision
Section heading	What is in the section	The heading names the layout and the reader skips it

Why do stale names cost more than bad ones?

We inherited a site nobody on our side had built. The project document listed a staging URL that had been correct a year earlier; by the time we arrived it served a different application that looked broadly similar. Production ran a different framework from the repository we had been handed. Publishing turned out to be a push to a branch rather than a deploy command, so the obvious command would have pushed a half-built version over the live site.

Nobody had lied. The name had outlived what it pointed at, and everyone downstream kept quoting it. We found it by making one trivial visible change and following it end to end before touching anything real.

What does a bad page title actually cost?

On one account the average organic click-through rate was 0.11% across a month while positions looked healthy. The audit pointed at titles and descriptions rather than rankings.

For a searcher the title is the whole page until they click. A title that names the topic asks the reader to guess whether the answer is inside.

How is a price a naming problem?

A price is the name for what a buyer owes. Pricing on one account had been structured per deal, and it confused sales conversations: prospects spent the call working out what they would owe rather than whether they wanted the product. It was replaced with flat tiers. The outcome was never measured, so the change is all I can report.

The same discipline runs one level up. On an account with one owner and three products, each brand needed its own voice and its own ICP. Shared ownership does not make one name work across three audiences.

A name that was accurate a year ago is more dangerous than an ugly one, because everybody downstream is still quoting it.

How do you check a name is still true?

Re-check names when you re-check access. Both go stale for the same reason: somebody changed something and nobody told the document everyone quotes.

For anything pointing at infrastructure, verify by doing rather than reading, and learn a client’s publish path before you have write access. For names a customer reads, the check is a stranger: if somebody outside the account cannot say what is behind it, it is a label.

The two engagements underneath this one are what a 0.11% click-through rate is telling you and verifying what is actually live.