Between a decent position in the results list and an enquiry in your inbox sits a single decision: someone skims three lines of text and clicks, or does not. Those three lines are the title link, the address line and the short description beneath it. They do not appear there by accident, and they do not necessarily come from what was once written into the page source. This article explains how titles and descriptions in search results come about, why search engines rewrite titles so often, which part stays visible in practice, and which wording patterns work for service pages, location pages and blog posts so that the click matches the page.
A result is made of three building blocks
An entry in the results list looks plain and is still assembled from parts. At the top sits the title link, the clickable headline. Above or below it runs a source line with the domain and often a path showing which area of the site the page belongs to. Beneath that comes the snippet, the two or three lines that are supposed to explain what awaits on the page. For a small business this trio is the real front door: it decides, before any layout, any photo and any loading time figure, whether anyone arrives at all. Competition for that moment is dense, because around 89 percent (Federal Statistical Office of Germany) of companies with internet access run their own website, and 93 percent (Federal Statistical Office of Germany) of people aged ten and over in Germany are online.
- Title link — the clickable line. It is usually built from the page's title element, but it can also be assembled from headings or other visible text.
- Source line — domain and path. It comes from the address structure and, where present, from a marked-up breadcrumb trail.
- Description in the snippet — the explanatory text. It is created primarily from page content; the stored meta description is used when it summarises the page more accurately (Google Search Central, Snippets in search results).
<head>
<title>Roof Repair in Leeds | Example Roofing</title>
<meta name="description"
content="Leaking roof or loose tiles? We repair pitched and flat roofs across the Leeds area. On-site survey, fixed price, appointment usually within two weeks.">
</head>Two entries, and nothing more is needed at this point. One distinction matters here: stars, opening hours, expandable questions or price ranges in a result come from an additional machine-readable layer. How that markup works is covered in the article on structured data and enhanced search results. This piece is only about the readable text: the headline people click and the sentence underneath that sets the expectation.
Why search engines rewrite titles
The stored title is a proposal, not an instruction. Google openly lists the sources a title link can be built from: the title element, the main visual title shown on the page, heading elements such as H1, the og:title meta tag, other large and prominent text, other text on the page, and anchor text in links pointing to the page (Google Search Central, Control your title links in search results). In practice, though, your own title almost always wins: after a revision of the title system, Google states that title elements are now used in around 87 percent (Google Search Central Blog) of cases, up from roughly 80 percent (Google Search Central Blog) before. Writing good titles means working with the system rather than against it.
The remaining cases are not random; they follow recognisable patterns. Google names four typical situations, drawn from the analysis of very large sets of pages. All four can be checked on your own website, and all four can be fixed with reasonable effort.
Half-empty titles
A template joins the page name and the business name, but the page name is missing. What remains is a title consisting of a separator and the company name (Google Search Central Blog).
Obsolete titles
A page is reused year after year, the visible headline names the current year and the title in the source still names the old one. The engine corrects it using the visible headline (Google Search Central Blog).
Inaccurate titles
The title promises content that only sometimes appears on the page, for instance in changing listings. The engine replaces it with wording that fits the actual content better (Google Search Central Blog).
Boilerplate in series
Several pages carry the same title because the distinguishing detail is missing. The engine adds it from large, visible heading text so the entries can be told apart (Google Search Central Blog).
A replaced title is a hint, not a penalty
Having a title at all is not the problem: 98.0 percent (HTTP Archive Web Almanac) of desktop pages and 98.2 percent (HTTP Archive Web Almanac) of mobile pages ship a title element. Quality is the actual building site — half-empty, copied and arbitrary titles are far more common than missing ones.
How much of it stays visible
There is no fixed character limit, and any number sold as a hard rule is a simplification. Google puts it plainly: title links and snippets are truncated as needed, typically to fit the device width (Google Search Central, Control your title links in search results). A narrow phone shows less than a wide screen, and the same page can appear trimmed differently for different queries. A planning figure is still worth having, so the core message reliably sits inside the visible area.
Reality on the web sits far above any sensible display length. The median title length is 12 words (HTTP Archive Web Almanac) and 77 characters (HTTP Archive Web Almanac) on desktop, or 79 characters (HTTP Archive Web Almanac) on mobile; in the top tenth it reaches 150 characters (HTTP Archive Web Almanac) and even 155 characters (HTTP Archive Web Almanac) on mobile. For descriptions the gap is wider still: the median is 40 words (HTTP Archive Web Almanac) and 272 characters (HTTP Archive Web Almanac). A large share of what businesses write therefore reaches nobody.
| Element | Planning figure | What happens when it runs long |
|---|---|---|
| Title link | around 55 to 60 characters, core message first | The end is cut off, usually the business name first |
| Description | around 150 to 160 characters, one complete thought | The rest falls away, or the engine assembles its own text |
| Town or region | within the first 40 characters when location matters | It slides out of the visible area and stops working |
| Business name | at the end, set apart by a separator | On truncation it pushes out exactly the statement that matters |
The first words carry the decision
The same order applies to the snippet text, only with more room. One complete, understandable sentence beats a string of keywords, because it is read rather than decoded. How to write sentences like that without slipping into advertising language is covered in the article on website copy customers actually read.
Fitting in location and service without a keyword chain
The most common mistake in local titles is stringing words together. "Roofer Leeds roof repair roof replacement Wakefield Leeds roofing" contains many words and no statement. Google advises against it explicitly: a few descriptive terms are sometimes helpful, but there is no reason to have the same words or phrases appear multiple times (Google Search Central, Control your title links in search results). There is a practical effect on top: a keyword chain reads like a directory entry and inspires no confidence.
Write descriptive and concise text for your title elements. Avoid vague descriptors like "Home" for your home page, or "Profile" for a specific person's profile.
In practice a simple order works: what is done, where it is done, who does it. "Roofer Leeds roof repair cheap company" becomes "Roof Repair in Leeds | Example Roofing". The town appears once, not three times. The word people actually search for sits at the front. And the business name takes the role it really has in a result: recognition. Which words customers actually use is the subject of the article on finding the search terms customers use.
- The service is named in the customer's language, not in the trade's internal language.
- The location appears exactly once, and only where it genuinely applies to that page.
- The business name sits at the end, set apart by a consistent separator.
- No word repeats, no hyphen chains, no strings of neighbouring towns.
- No absolute promises such as "best" or "cheapest" — legally risky and ineffective in the result.
- The title matches what the page actually offers, not what the business would like to offer.
With several locations, the question shifts from wording to structure. Every location needs its own page with its own title and description; a combined title listing five towns helps neither people nor machines. How local visibility is built up cleanly is described in the article on visibility for local businesses.
Every page gets its own text
The meta description is the most frequently copied text on a website. That is precisely what devalues it. Google states the expectation clearly: identical or similar descriptions on every page of a site are not helpful when individual pages appear in search results — each page needs its own description (Google Search Central, Snippets in search results). Adoption shows how much room is left: only 66.7 percent (HTTP Archive Web Almanac) of desktop pages and 66.4 percent (HTTP Archive Web Almanac) of mobile pages ship a description at all.
At the same time, nobody should expect the stored text to appear verbatim. In a study of 20,000 search terms (Ahrefs), the meta description was not used in 62.78 percent (Ahrefs) of cases and was replaced by text generated from the page content instead. For very broad queries the share was 59.65 percent (Ahrefs), for specific longer queries 65.62 percent (Ahrefs). And 25.02 percent (Ahrefs) of well-ranking pages had no description stored at all.
A copy template is not a description
Browser makers' auditing tools take the same view. The automated check for descriptions fires when the element is missing or its content is empty, and it recommends a unique description per page, clear and concise, avoiding vague wording such as "Home" (Google, Lighthouse documentation). The developer documentation adds the editorial expectation: the description should be a short and accurate summary of the page's content (web.dev). Multilingual sites add another layer, because titles and descriptions belong to each language version in their own right — the connections are covered in the article on multilingual websites and hreflang.
Finding duplicate and empty titles in what you have
For a website with fifteen to eighty pages, the audit needs no specialist tool — a spreadsheet and an hour will do. What matters is that every page sits in one list and can be read side by side; duplicates only become obvious once they are stacked underneath each other.
- List all published pages, most easily from the sitemap or from your system's page overview.
- Transfer the title and description of every page into a table, along with the character count of both fields.
- Sort by title: anything appearing identically more than once is a candidate for a rewrite by the search engine.
- Mark empty fields — missing descriptions and titles consisting only of the business name come first.
- Check conspicuous lengths: anything well beyond 60 characters in the title and 160 characters in the description.
- Finally, read each entry against the page itself: does the title state what the page actually answers?
The Search Console performance report then provides the counter-check. There, impressions are the number of times the site appeared in results, clicks are the visits that followed, and the click-through rate is simply the click count divided by the impression count (Google Search Console Help, Performance report). Pages with many impressions and few clicks are the most interesting candidates, because visibility already exists and only the text in the result fails to convince.
The audit pays off most before a rebuild. Merging pages or changing addresses moves titles and descriptions along with them — the groundwork required is described in the article on a website relaunch without losing rankings.
Wording patterns for the main page types
Patterns are starting points, not stencils. They prevent the two most common mistakes — the wrong thing sitting at the front, and every page sounding the same. The overview below covers the page types found on almost every business website.
| Page type | Pattern for the title | Pattern for the description |
|---|---|---|
| Home page | Main service + town + business name | What the business does, for whom, in which area, plus the next step |
| Service page | Service + town + business name | The trigger in the customer's language, then process and conditions in one sentence |
| Location page | Town + service + benefit | Reachability, service area and a concrete note on appointments or response time |
| About | Short claim + business name | Who is behind it, since when, with which qualification — without self-praise |
| Contact | Contact + town + business name | Which channels exist, when someone is available and what an enquiry should contain |
| Blog post | Question or benefit + business name | What the text answers and who it is written for, in one complete sentence |
Two notes on using the patterns. First, the business name may be dropped when the title would otherwise run long — the domain already appears in the source line. Second, not every page needs the same structure. A pricing page tolerates a different tone than an emergency callout page. Which pages are worth having at all is covered in the overview of the pages a business website needs.
Sales copy or clear expectations
There is a temptation that works in the short term and hurts in the long run: tuning the title for maximum curiosity. "What you need to know about your roof" may generate clicks, but somebody who needs a repair is gone again after two seconds. Those bounces help nobody. The business pays for them with attention it needed elsewhere, and the searcher pays with time.
The better yardstick is clarity of expectation: the title describes what the page delivers, the description adds the conditions. "Roof repair in Leeds" plus "on-site survey, fixed price, appointment usually within two weeks" filters deliberately. Someone with a flat roof in another region does not click — and that is a gain, not a loss. Enquiries that match the offer are the actual purpose of the page.
There is a legal side as well. Advertising superlatives in a title are not harmless under competition law: claims to a leading position and absolute promises must be substantiated, otherwise they are open to challenge. Wording such as "unbeatably cheap" or "the best firm in the region" therefore belongs neither in the title nor in the description. Factual details — service, area, process, availability — are not only safer, they are the better sales arguments. How to hold that tone across an entire website is shown in the article on website copy that gets understood.
Measuring instead of guessing
Whether a new title performs better cannot be read from a gut feeling, but from the click-through rate in relation to position. A jump from two to four percent at an unchanged average position is a signal; the same jump alongside an improved position is not, because the position itself lifts the rate. The Search Console reports both figures next to each other, which is exactly why they should be read together (Google Search Console Help, Performance report).
- Do not change everything at once: five to ten pages per round is enough, otherwise the effect cannot be attributed.
- Note the figures before the change — impressions, clicks, click-through rate and average position for the same period.
- Use an equally long comparison period and account for seasonality; a December is no yardstick for a June.
- Always look at position and click-through rate together, otherwise a ranking gain is booked as a copy gain.
- Plan for patience: changes to titles and descriptions only take effect after the page is recrawled and reprocessed, which can take a few days to a few weeks (Google Search Central, Control your title links in search results).
If the search engine keeps showing its own title despite clean work, that is no reason for hectic activity. The sensible next look is at the page's visible main heading: it is one of the sources an alternative title link is built from. When the title element and the H1 agree in substance, the most common trigger for a rewrite disappears on its own.
How XICflow handles titles and descriptions
In many systems, title and description are optional fields buried in the third tab of the page settings. That is exactly why they stay empty or get copied. In XICflow both fields belong to the page itself: a page only counts as complete once title and description are filled in, and both fields show while typing how much of them is likely to remain visible in a result.
Required field with length check
Title and description are separate fields per page and per language. A length indicator flags entries that are too short, too long or empty before the page goes live.
Suggestion, not automation
From content, page type and location context the assistant generates suggestions for title and description. Nothing is applied on its own — every suggestion passes through editorial approval.
Overview across all pages
One list shows the titles and descriptions of every page side by side. Duplicate, empty and overlong entries become visible without exporting a spreadsheet.
Because every page in XICflow exists as a structured configuration, these fields do not depend on a template that gets lost in the next rebuild. How content, structure and delivery work together is shown in the product feature overview and in the walkthrough of how a website is built. If you would rather look than read, there are finished examples among the demo websites; the terms are set out in the plan overview with AI credits, and the comparison with other routes to a website is covered under why XICflow.
Sources and studies