Writing for the search result, not just the page

A page competes twice. First as two lines of text in a list of ten similar-looking results, where the only question is whether anyone clicks. Second as the page itself, where the question is whether they stay. Almost all the effort goes into the second competition, and it does not begin until the first one is won.

The two lines that do the work

The title and description shown in a search result are the advert for the page. They are frequently written last, by whoever was closing the task, and they show it.

One point worth internalising: search engines rewrite titles they judge unhelpful, and a rewritten title is often worse than what you would have written honestly. The reliable way to keep your own title is to make it accurately describe what the page delivers, which is also the way to avoid the visitor leaving immediately when it does not.

Headlines on the page are a different job

The headline on the page has a different task from the title in the search result. The search title has to earn a click from someone comparing options. The page headline has to confirm, within a second of arrival, that the visitor is in the right place.

That means the two can differ, and often should. But they cannot contradict, because a visitor who clicked one promise and landed on a different one leaves immediately — and that immediate departure is itself a signal that the result was not useful.

How long, and how hard to read

Two measurements are worth taking and both are commonly misused.

Length has no correct value. There is no threshold that ranks better, despite a persistent industry belief otherwise. What length is useful for is comparison: if every page answering a question runs to a thousand words and yours runs to two hundred, you have probably not answered it. The number is a prompt to check, not a target to hit — and writing to hit it produces exactly the padded article everyone recognises.

Reading level is more useful and more often resisted, because it sounds like an instruction to simplify. It is not. It measures sentence length and word complexity, and what it actually detects is the writing that requires a second pass to parse. A specialist reader in a hurry benefits from clear sentences exactly as much as anyone else; the expertise should be in what you say, not in how much work it takes to read it.

Keyword density, and why it survives

Keyword density as a target — aim for two per cent — is genuinely obsolete, and writing toward it produces text that reads as though it were written toward it. That is the whole problem: the pattern is detectable by readers, which makes it detectable by everyone.

The measurement is still worth taking, as a diagnostic rather than a goal. Unusually high repetition of one phrase usually means a real writing problem — a paragraph circling without progressing, or a term used where a pronoun would do. Fixing that makes the text better, and the density falls as a side effect rather than as an objective.

The order to work in

1. Write the page for the person who needs the answer, without thinking about search at all.
2. Check reading level and fix anything requiring a second pass to understand.
3. Compare length against what else answers this question, and check for gaps rather than adding words.
4. Scan for accidental repetition, and fix the writing rather than the count.
5. Write the page headline so a visitor knows within a second they are in the right place.
6. Write the search title and description last, deliberately, and preview them next to the competition.

The last step being last is the point. Written first, it constrains the page toward the phrases you hoped to use. Written last, it describes what the page genuinely turned out to be, which is both more honest and more likely to survive contact with a search engine that rewrites overclaiming titles.