The words your customers actually use

Somewhere in your product there is a feature with a name you invented. Maybe it is capitalized. Maybe it has a small icon next to it in the navigation. You have said it out loud so many times — in the demo, on the pricing page, to your own family at dinner — that it has stopped sounding like a made-up word and started sounding like the obvious word. It is not the obvious word. Nobody outside your company has ever typed it into a search box, and quite a few of the people who would happily pay you are, at this moment, typing something else entirely and finding somebody who used their phrasing instead.

This is one of the cheapest problems in marketing to fix and one of the easiest to never notice, because the feedback is silent. There is no report that says forty people looked for this today and you were not there. There is just a number that is lower than it should be, spread thin across every channel at once, with nothing obviously wrong anywhere.

The same thing, three vocabularies

The trap is that your naming is not arbitrary. It is the result of real thought, usually about the mechanism — what the feature does internally, how it differs from the thing you built before it, what the category will eventually be called once everyone catches up. All of that is coherent. It is also the vocabulary of someone who already understands the problem deeply.

Your buyer is not there yet. They do not have a name for your category, and they are not looking for one. They have a bad afternoon and a phrase for it, and the phrase is almost always more literal and less elegant than anything you would put on a homepage. They do not search for unified attribution surface. They search for where my signups come from. They do not want lifecycle orchestration. They want to stop forgetting to email people who signed up last week.

people do not search for your category. they search for their bad afternoon.

Two different words for the same feature are not a style preference, then. They are two different populations of people, and only one of them is looking for you.

Where the words are already sitting

You do not have to guess at this, and you should not run a survey about it either. The words already exist, written down, in places you can read this afternoon.

Start with the search queries that already bring people to you. Search Console shows what people actually typed before they landed on your site, and the interesting rows are never the ones you optimized for — they are the ones where somebody found you by accident, using a phrase that never appears anywhere on your site. Those rows are a gift: real demand, in the customer's own words, arriving despite your copy rather than because of it.

Then read your own inbox. Support questions, sales replies, the first message a new customer sends. When someone describes what they were trying to do in their own words, copy that sentence somewhere. Do the same with the notes you take when a signup tells you what brought them in — the same raw material we talked about in asking people how they found you, used for a different purpose. Within a couple of weeks you will have a small file of phrases, and the repeats in it are your headline.

What it is worth, roughly

Invent some numbers, because the shape is what matters. Say a phrase your customers use gets looked up three hundred times a month, and you currently rank nowhere for it because the phrase appears zero times on your site. Suppose that publishing one honest page built around that phrasing eventually earns you a modest position — call it thirty visits a month, of which four percent sign up. That is a bit over one signup a month from one page, which sounds unimpressive until you notice it repeats every month without further spend, and that you probably have five or six such phrases rather than one.

The same swap costs nothing at all in the places you are already paying for. An ad that uses the reader's phrasing and a landing page that repeats it back are not a new campaign, just different words in an existing one — and the click-through difference between your word and their word can be larger than most of the bid tuning people agonize over. This is the same match problem as when your ad and your page disagree, one step earlier: before the ad can agree with the page, both have to agree with the customer.

The part that is genuinely hard

Here is the tradeoff, stated plainly, because people skip it and then get annoyed. Customer language is usually flatter and less distinctive than the language you invented. If you write every page in the most searched-for phrasing, you end up sounding exactly like everyone else in your category, which solves the finding problem and creates a choosing problem.

So do not replace your vocabulary. Lead with theirs and land on yours. The headline, the ad, the page title, the email subject line — the places where someone decides in half a second whether this is about them — use the words they already have. Then, once they are reading, introduce your name for it, because a distinctive name is genuinely useful to someone who has decided they need the thing and now wants something to call it. Invented words are good closers and terrible openers.

This is also the part where drafting with AI helps and quietly hurts. When we draft copy for a page or an ad, the draft is fluent by default, and fluent writing gravitates toward the polished, generic, category-level phrasing — it sounds like marketing because it was trained on marketing. That is exactly the register you are trying to escape. Which is why the draft waits for you rather than going out on its own: the one edit that matters most is usually swapping the smooth phrase for the clumsy real one somebody actually said to you in an email last Tuesday.

A rule of thumb to end on: if a phrase on your homepage could not plausibly be typed, word for word, by a person who has never heard of you — it belongs further down the page.