On-Page SEO Basics for Guide Sites

The title element on your trip page is one of nine things Google looks at when deciding what to print as the title of your search result. If it does not like what it finds, it will use the anchor text from somebody else's website instead.

That is not speculation. Google lists all nine sources, names seven specific defects that cause it to override a title, and states that the generation of these titles is completely automated. Meanwhile the HTML standard, which defines the element in the first place, says something Google's guidance assumes but never spells out: the title has to make sense to a person who encounters it with no other context at all. Between those two documents and one internet standard about addresses, the whole of on-page work for a guide's site is settled. Everything that sits beside this is filed under the getting-booked hub.

What is the title element for?

Identifying the page when nothing else is around it.

The HTML standard says the title element represents the document's title or name.

It then gives the instruction that matters: authors should use titles that identify their documents even when they are used out of context, giving as examples a user's history, their bookmarks, or search results.

Out of context is the operative phrase, and a search result is precisely that situation.

A page titled Trips, or Rates, or Booking, identifies nothing once it leaves the site it belongs to.

The same standard specifies that there must be no more than one title element per document, that it lives in the head, that neither of its tags may be omitted, and that its content is text rather than markup.

None of that is advice. It is the definition of the element every search engine reads.

The specification is published as the HTML Living Standard.

Fishing With Captain Kirk, WI

Why is the title different from the heading?

Because the heading gets to assume you are already there.

The standard states directly that a document's title is often different from its first heading, since the first heading does not have to stand alone when taken out of context.

Its worked example is a book chapter titled Introduction to The Mating Rituals of Bees whose first heading is simply Introduction.

A second page carries the title Dances used during bee mating rituals with the heading The Dances, and the standard notes the title describes the subject matter unambiguously while the heading assumes the reader knows the context.

Translate that to a guide's site and the pattern is immediate.

The heading on your half-day page can read Half day, because the visitor arrived through your own navigation and knows whose half day it is.

The title cannot, because it is going to appear beside eight strangers in a list.

The pages that need this treatment most are described in the water and city pages piece.

Is your title what Google prints?

Sometimes, and it is one of nine inputs.

Google defines a title link as the title of a search result that links to the web page, and states that its generation is completely automated, taking into account both the content of a page and references to it that appear on the web.

The listed sources are content in title elements, the main visual title shown on the page, heading elements such as h1, content in og:title meta tags, other content that is large and prominent through style treatments, other text contained in the page, anchor text on the page, text within links that point to the page, and WebSite structured data.

Note the seventh and eighth of those. Text within links pointing to your page is somebody else's writing, on somebody else's site, and it is an input.

Google adds that it cannot manually change title links for individual sites, and that noticing updates to any of these sources requires a recrawl and reprocess, which may take a few days to a few weeks.

So a title change is not an edit. It is a submission.

The guidance is published as Google's documentation on title links.

What makes Google override a title?

Seven named defects, and it publishes all of them.

Half-empty title elements, where part of the text is missing, leaving something like a bare pipe followed by the site name.

Obsolete titles, where a page reused year after year still carries last year's date while the visible heading says otherwise.

Inaccurate titles, where the element does not reflect what the page is actually about.

Micro-boilerplate text, where a subset of pages share a title that omits the thing distinguishing them.

No clear main title, where more than one heading carries the same weight and prominence.

A mismatch of writing system or language between the title and the page's primary content.

And duplication of the site name, where Google may simply omit it because the result already shows it.

Each of those is a case where Google says it may generate a replacement from anchors, on-page text or the other listed sources.

Which of those catches guides?

Micro-boilerplate, almost every time.

The published example is a television site whose season pages all share a title that omits the season number, so it is unclear which page is which.

Google's fix in that case is to detect the season number in large prominent text and insert it into the title link.

The guide equivalent is a set of water pages or town pages that all read the same way with one word swapped, or worse, that do not swap it at all.

The guidance is direct about the underlying rule: have distinct text that describes the content of each page, and avoid long titles varying by only a single piece of information.

Its own example of the failure is a title reading like a band name followed by a list of everything the site offers, which it describes as containing a lot of uninformative text.

Its suggested remedy is to update the title dynamically to reflect what the page actually holds, including a word only when that page contains the thing.

Writing pages that genuinely differ is the subject of the blogging piece.

How long should the title be?

There is no limit, and it gets cut anyway.

The guidance says plainly that there is no limit on how long a title element can be, and that the title link is truncated in search results as needed, typically to fit the device width.

Typically to fit the device width is the useful half of that sentence, because it means the cut-off is not a number you can design against.

What follows practically is to put the distinguishing words first and the site name last.

Google's own suggestion is to brand titles concisely, including just the site name at the beginning or end, separated by a delimiter such as a hyphen, colon or pipe.

It also warns against vague descriptors such as Home for a home page or Profile for a person's profile, and against keyword stuffing, giving the example of a title repeating the same phrase in four forms.

A guide's version of that failure usually involves a species, a town and the word charter appearing three times each.

Which phrases clients genuinely type gets its own inventory in the keyword piece.

No-Fish No-Fee, MI

What is a URL, in standards terms?

A hierarchy, and the parts have different rules.

The internet standard that defines URIs describes the path component as data, usually organised in hierarchical form, which along with the query component identifies a resource within the scope of the scheme and naming authority.

It specifies that the path is terminated by the first question mark or number sign character, or by the end of the URI.

It also requires that where a URI contains an authority component, the path must either be empty or begin with a slash.

None of which sounds like marketing, and all of which decides whether two addresses on your site are the same address.

That question matters more than it looks, because a search engine that treats two versions as different pages is splitting whatever either of them has earned.

The specification is published as RFC 3986.

Why does capitalisation matter in a path?

Because only part of the address is case-insensitive.

The specification's normalisation rules state that the scheme and host are case-insensitive and should be normalised to lowercase, giving the example that an address written with capitals in the domain is equivalent to the same address in lowercase.

It then says the other generic syntax components are assumed to be case-sensitive unless a scheme specifically defines otherwise.

The path is one of those other components.

So a page at a path with a capital letter and the same page at a lowercase path are, by the standard, two different addresses.

Whether your server treats them as the same is a configuration question, and different servers answer it differently.

The practical instruction is to pick lowercase paths, use them everywhere, and link internally in exactly that form.

The wider foundations behind this are set out in the search basics piece.

What if the page cannot be crawled?

Your title stops mattering and somebody else's writing takes over.

The guidance warns about disallowing search engines from crawling pages, noting that the robots protocol can stop Google crawling a page but may not always prevent it being indexed.

It gives the mechanism: Google may index a page discovered by following a link from somebody else's site.

Then the consequence, stated plainly: if Google does not have access to the content on the page, it will rely on off-page content to generate the title link, such as anchor text from other sites.

So a blocked page can still appear, titled by whatever a stranger typed into a link.

To keep a URL out of the index the guidance points instead to the noindex rule, which is a different mechanism entirely and requires the page to be crawlable to be seen.

Confusing those two is one of the more expensive small mistakes available on a small site.

How links from elsewhere accumulate in the first place is covered in the backlinks piece.

What should a guide's page actually carry?

Six things, and none of them is a plugin.

One title element, in the head, unique to the page, that identifies it with nothing else around it.

One clearly dominant heading, since Google names no clear main title as a defect and suggests the main title be the first visible h1 and visually distinct.

A path in lowercase that matches how you link to it everywhere else.

Body text in the same language and writing system as the title, which Google names as a mismatch defect when they differ.

Enough genuinely distinct content that the page is not a sibling of seven others with one word changed.

And the site name handled once, concisely, at one end of the title rather than repeated through it.

The structured data that sits underneath is covered in the schema markup piece.

Where do guide sites go wrong?

Seven ways, and the first three are the same failure.

Titling pages Home, Trips, Rates and Contact, which identify nothing out of context.

Running one title template across every page with a single word varying, which is the micro-boilerplate case by name.

Repeating the full business name and slogan in every title, which the guidance describes as looking repetitive when several of your pages return for one query.

Stuffing the same species and town into the title several times over, which the guidance calls out directly.

Using mixed capitalisation in paths, then linking to both forms internally.

Giving three headings identical visual weight so no main title is apparent.

And blocking a page in the robots file to keep it private, which may leave it indexed under a title written by a stranger.

Setting up the Business Profile alongside all this belongs to its own piece.

What do site owners not expect?

That the title is a suggestion rather than an instruction.

Most people believe the title element is what appears, and it is one of nine inputs into an automated decision.

The second surprise is that anchor text on other people's sites is one of those inputs, which means your result can be titled by a stranger's phrasing.

The third is that the HTML standard has been telling authors to write out-of-context titles since long before anybody optimised anything.

The fourth is that paths are case-sensitive by the standard, which most people discover through a duplicate-content problem rather than by reading.

The fifth is that a title change takes a recrawl and reprocess to register, which the guidance says may take a few days to a few weeks.

Which means judging a title change after a weekend tells you nothing at all.

A saner approach to watching that is described in the rank tracking piece.

How would you audit your own pages?

Five checks, an hour, using nothing but a browser.

List every page and read its title element as though you had never seen the site, asking whether it identifies the page alone.

Put those titles in a column and look for the ones that differ by a single word, which is the micro-boilerplate pattern.

Check each page has exactly one dominant heading rather than three that look the same.

Type each path in mixed case and see whether your server serves the page, redirects, or errors, since the standard treats the forms as different.

Then search your own business name and compare what Google prints against what you wrote, because the difference is the diagnosis.

Anything Google is rewriting is a page it has judged against one of its seven named defects.

Which listings beyond your own site matter is assessed in the directories piece.

On-page, in order

One title per page, written for a stranger, then leave it.

Write each title so it identifies the page in a list of results from sites the reader has never visited.

Keep the heading free to be shorter, because the visitor is already there and the standard expects the difference.

Put the distinguishing words first and the site name once, at one end, behind a plain delimiter.

Make sure no two titles differ by a single word, and rewrite the ones that do.

Use lowercase paths, everywhere, including in your own internal links.

Never use the robots file as a privacy measure on a page you would mind being titled by a stranger.

Then wait weeks rather than days before judging anything, because the documentation says that is how long it takes to register.

Layering local search over the top belongs to its own piece, and the map results to the map pack piece.