Title tags: how to write one that gets clicked, and why Google rewrites yours

In an audit of a 312-page site, rewriting titles was the most profitable action in the whole project. Not new content, not links: the titles. One page sat at position 1 with 124 impressions and zero clicks because the title Google showed did not contain the term people were typing. It is the only SEO work that costs minutes per page and can change the result the same month, and almost no plan does it before writing article fifty-one.

How do you write a title tag that works?

A good title tag contains the term people type, reads as if a person wrote it, and describes what is inside. You measure it by the impressions and clicks of the same query, not by position, because a page can rise in rank and still lose clicks when its title does not match the search.

Quick answer

• It is the highest-return edit on most sites, and it takes minutes.
• Google rewrites the title in a share of cases, and knowing why saves you fighting it.
• The term people type has to be inside it. That is half the work.
• You measure it with impressions and clicks on the same query, not with position.

What kind of work this is

The framing that wastes time is treating it as copywriting. Under that idea the goal is a catchy title, and catchy competes with comprehensible.

What decides a click on a results page is not wit, it is recognition. The person scanned ten titles in two seconds looking for the one that contains what they just typed. A title that does not contain it is discarded before it is read in full, however good it is, and the same recognition decides whether a generative engine treats the page as a citable answer. From that comes the rule that sums up the rest: the title has to contain the search and sound like a person. Both at once, and in that order of priority. Which term to put in, when the page does not exist yet, is found the way any query is, in free keyword research.

Why Google rewrites the one you set

It is one of the things that most confuses people and it has an explanation. Google's own documentation on title links, opened on 16 September 2026, says its generation "is completely automated and takes into account both the content of a page and references to it that appear on the web", and lists the sources it draws from: the <title> element, the main visual title, heading elements such as <h1>, the og:title tag, prominent text, anchor text, and links pointing to the page.

It reaches for one of those instead of your title in a few situations the same page names, and they are avoidable. Half-empty titles, where part of the text is missing. Obsolete titles, "when the same page is used year-after-year for recurring information, but the <title> element didn't get updated." Keyword-stuffed titles like "Foobar, foo bar, foobars, foo bars" that "don't help the user". And titles that do not accurately reflect the page. On length, Google is explicit: "there's no limit on how long a <title> element can be", but the title link "is truncated in Search results as needed, typically to fit the device width."

The practical consequence is that fighting the rewrite does not work, and it does not need to: a title of the right length, without stuffing, faithful to what is inside, is kept most of the time. And if Google insists on using the H1, that is usually a sign the H1 is better than the title you wrote.

How to write it, in order

First, the query. Open the Search Console performance report, filter by that page, and see what the people who see it are typing. That term goes inside. The reason it goes first, and not the brand, is that recognition beats everything else on the results page.

Then, what the page delivers. A title that promises more than the page holds buys a click and loses the person, and that worsens the signal instead of improving it. What the click leads to, and how intent is read off the results page, is in search intent.

Then, the length. The limit is pixels, not characters, so the exact figure varies with the letters you use. As a practical reference, around sixty characters is where the risk of truncation begins, and what matters is that the part describing the page fits before the cut, not that the whole title fits.

And last, the number or the specific detail, if they are true. "12 causes" or "with a seven-month curve" work because they promise something verifiable. Inventing the number to make it work is a bad trade: you win a click and lose the reader, the same trade weighed for the whole page in how to write SEO content.

The title against the H1, and against the meta description

Three fields get confused. The title tag is what shows as the clickable line in results and in the browser tab; the H1 is the visible headline on the page; the meta description is the grey text under the title, which is not a ranking factor and which Google also rewrites often. The title and the H1 can differ, and often should: the title carries the term as people type it into a search, the H1 reads for someone already on the page. When they are identical across a whole site it is usually because both were generated, not written.

When NOT to touch the title

When the page ranks in the top few and converts. Google already validated it; rewriting the title risks a position that works to fix something that is not broken, the same read as a ranking drop in why organic traffic drops.

When the CTR is already good and impressions are low. That is a position problem, not a title one, and a new title does nothing for it.

When the page targets a query it should not. A better title for the wrong intent still brings the wrong reader.

Mistakes that repeat

• Writing the title as an ad. Recognition beats wit; the term people type has to be in it.
• Leading with the brand name. On a site with no fame it takes the space the query needed.
• Chaining terms with pipes. The fastest way to get the title rewritten, and it reads as what it is.
• Measuring the title by position. A page can rise and lose clicks; impressions and clicks on the query are the measure.
• Inventing the number in the title. It wins a click and loses the reader when the page does not deliver it.
• The same title on the whole site. Identical titles are the tell of generation, not writing.

Data and transparency

That Google's title links are generated automatically from the title element, the main visual title, headings, og:title, prominent text, anchor text and links to the page, that it rewrites for half-empty, obsolete, stuffed or inaccurate titles, and that there is no character limit while the link is truncated to the device width, is quoted from Google's documentation on influencing title links, opened on 16 September 2026. That the meta description is not a ranking factor is Google's long-stated position. The 312-page audit with a page at position 1 and zero clicks for a title that did not contain the searched term is own client work, not named, with rounded figures. Search demand for "title tag", about 2,400 monthly searches in the United States with a keyword difficulty of 39, comes from Ahrefs on 16 September 2026. The pixel-length reference is trade judgment, not a published threshold. Verified as of September 2026.

Primary source, opened 16 September 2026: Google on influencing your title links.

What this changes

The title is treated as the last cosmetic step, written in the final minute before publishing and rarely touched again. On most sites it is the opposite: the single edit with the highest return and the shortest turnaround, because it is the only one Google can reward within the same month.

What changes when you see that is the order. The title stops being decoration on top of the page and becomes the first thing checked on any page that gets impressions and no clicks, before a word of the content is rewritten. The page was already being shown. It was just being shown under a name the searcher did not recognize as theirs.

Frequently asked questions