Meta Tags and Site Structure: A Practical Guide
A practical guide for small business website owners | Reviewed 16 October 2026
Metadata describes a page or gives supporting systems instructions about it. Site structure is how your pages are organised and linked. Together, they help people understand what a page offers and find the information that belongs with it.
Start with clear page purposes, accurate titles and useful navigation. Treat indexing controls separately: a noindex setting or a canonical change can have a very different effect from editing a description.
This guide explains the important distinctions and shows how to review them in WordPress. For a complete individual-page workflow, use the on-page SEO guide; for access and indexing faults, use technical SEO.

Know which element you are editing
SEO plugins often group these settings in one panel, but they do different jobs. A title element and a canonical link are not technically meta tags, even though they are often discussed alongside them.
| Element | Main purpose | Important distinction |
| HTML title | Describe the page in the browser and provide a source for search title links. | The title you supply is not guaranteed to be the title Google displays. |
| Meta description | Summarise the page as a possible search snippet. | Google can select different text from the page. |
| Robots metadata | Give supporting crawlers indexing or presentation instructions. | Not a privacy control or a ranking boost. |
| Canonical link | Signal a representative URL for duplicate or similar content. | A preference signal, not a redirect or a command. |
| Open Graph metadata | Describe a page for supported sharing and other systems. | Display varies by platform; some properties can also inform Google. |
| Visible heading | Identify and organise content on the page. | An H1 is part of the page body, not hidden metadata. |
Write titles that identify the actual page
Use a distinct, concise HTML title that describes the subject. Include a useful distinguishing detail such as a real location, product or service. Avoid repeated keyword strings or using “Home” and “Our Services” where a clearer description is available.
A fictional upholstery cleaner might use “Sofa Cleaning in York | Willow Cleaning”. That would only be appropriate if the business and page genuinely offer that service in York.
Google automatically generates title links using several possible sources, including the HTML title, main visual heading, other page text and references to the page. Its documentation also lists og:title. Keep these descriptions consistent rather than assuming one field overrides everything else.
There is no universal 60-character limit for an HTML title. Search display can be shortened to fit the available space. A plugin preview helps you judge clarity, but does not guarantee the exact result.
Google’s guidance on title links
Use descriptions to set accurate expectations
A meta description is a short summary Google may use beneath a search result. Google can instead select text from the page, and the snippet may vary with the question.
Explain what the reader will find. For the fictional cleaner: “See what our York sofa-cleaning service includes, which fabrics need assessment and how to request a quote.” The page must actually provide those details.
Do not copy the same description across unrelated services or invent an offer to make it more persuasive. Keep the wording useful and concise, but do not treat 160 characters as a Google rule. Snippet length depends on the display.
A description is not a direct ranking boost, and more clicks do not automatically mean more suitable customers. Evaluate whether it attracts the right visitor, alongside the page and wider search context.
Google’s guidance on snippets and descriptions
Handle robots settings as instructions
Robots metadata can be supplied in the HTML head or through an HTTP response header. It can affect indexing and how content is presented by search engines that support the instruction.
Index and follow are default behaviours for Google; adding those words is not a guarantee of indexing. A noindex instruction prevents inclusion when Google can crawl the resource and read the rule. Blocking access in robots.txt can stop it seeing that instruction.
Do not use noindex as a way to protect private information. Use proper access restrictions for private content. Nor should you apply noindex to every duplicate URL: canonicalisation may be the relevant solution, depending on what the pages are for.
Other robots instructions can restrict snippets or previews. Do not apply a restrictive site-wide template simply because it appears in an SEO checklist. Record what you intend to achieve and confirm its effects before changing it.
Google’s robots metadata specifications
In WordPress, inspect the relevant page setting, site-wide template and actual output if something unexpected appears. A setting in one plugin may not explain a conflicting server header or a second plugin’s output.
Keep canonical preferences consistent
A canonical link signals your preferred version of duplicate or very similar content. Google can select a different canonical. The annotation does not send a visitor to another page as a redirect does.
Keep internal links, sitemap entries and canonical preferences aligned. Investigate conflicting output before adding another tag. Do not point unrelated service pages at the homepage to “tidy” the index.
If the problem involves product filters, language versions or a site move, follow the relevant technical guidance. Do not remove a useful page from search merely to clear a duplicate warning.
Google’s canonical URL guidance
Understand sharing metadata and the fields you can ignore
Open Graph properties such as og:title, og:description, og:image and og:url describe a page for systems using the protocol. Platforms may use them to build a preview, but their presentation and caching behaviour differ.
Set a relevant image and description, then check the actual shared result where it matters to your audience. Keep the offer consistent with the page. Do not promise that a sharing image or title will appear identically everywhere.
Avoid the blanket claim that these properties are invisible to Google. Google documents og:title among its title-link sources. That does not mean every social tag controls rankings or that changing a preview guarantees more clicks.
You can leave the meta keywords field out of your Google SEO work: Google says it does not use that tag for indexing or ranking. A viewport setting concerns mobile presentation; it is not a place to put target phrases.
Meta tags and attributes Google supports
Organise pages around what customers need
A useful site structure gives visitors a clear route to the right information. Group related services, products or guides under meaningful labels, and use links to show the relationships between them.
Begin with a short inventory of important pages and the tasks they support. Look for unclear groups, near-identical pages and important resources that are hard to reach. The answer may be a better menu or contextual link, rather than a redesign.
A parent page should help people choose among its supporting pages. It should do more than repeat their titles. An advice guide can answer a different question and link to the relevant service where it helps the reader.
Content planning for customer questions and AI search
An example for a local service business
The following structure is fictional. The page names illustrate information roles, not a requirement to create all these pages.
| Page or group | Reader task | Useful connection |
| Services overview | Choose the relevant cleaning service. | Links to the real services with brief explanations. |
| Sofa cleaning | Check suitability, scope and the quote process. | Links to fabric-care guidance and the enquiry route. |
| Carpet cleaning | Understand this distinct service. | Links to relevant preparation information. |
| Areas served | Check whether the business covers the location. | Links back to relevant services and contact details. |
| Care guides | Find practical advice for a particular problem. | Links to supporting guides or services when useful. |
The business can improve those relationships without changing every URL into a new folder structure. Navigation hierarchy and URL folders are related design choices, but they do not have to mirror each other exactly.
Use internal links rather than a rigid click rule
Important pages should have useful links from other relevant pages. Link text should help someone predict the destination, such as “check our collection areas”. Use ordinary crawlable links rather than relying only on script-triggered controls.
There is no universal requirement for every page to sit within three clicks of the homepage. A small service site and a large product catalogue need different structures. Keep important journeys short and understandable without flattening everything into one enormous menu.
An orphan page has no useful internal route from the rest of the site. It may still be discovered through a sitemap or external link, but readers browsing your site can miss it. Add a relevant connection, or reconsider whether the page belongs in the site.
For shops, connect categories, subcategories and products in a way people can browse. Google’s ecommerce guidance explains that links help it understand the relative importance and relationships of pages; URL shape alone does not communicate the whole structure.
Google’s ecommerce site-structure guidance
Use categories and breadcrumbs with a purpose
WordPress categories and tags can generate archive pages. Keep them when they provide a useful way to browse. Avoid creating a new tag for every wording of a keyword or generating many nearly empty archives.
Do not apply a blanket index or noindex setting to all archives without assessing them. A carefully maintained category can be an important destination; another archive may add little for readers. Confirm the intended role before changing its search eligibility.
Breadcrumb navigation shows a page’s place in a hierarchy, rather than necessarily reproducing the visitor’s browsing history. It can help people move to a broader section. Check that the labels and linked destinations reflect the structure you intended.
BreadcrumbList structured data can support Google’s breadcrumb feature, subject to requirements. Google currently documents that feature for desktop Search. Useful on-site breadcrumbs can still serve readers on mobile; neither markup nor validation guarantees a particular search display.
Google’s breadcrumb documentation
Review metadata and structure together in WordPress
Choose a few important pages and use the following sequence. Check more than one template if the site includes services, products and posts.
- Record each page’s URL, reader task and intended search eligibility.
- Check the visible title and headings, then the title and description supplied by the SEO plugin.
- Confirm that templates do not produce repeated titles or descriptions across unrelated pages.
- Inspect the output for conflicting robots or canonical instructions when there is a problem.
- Follow the menu, breadcrumbs and relevant internal links on a phone.
- Check the sharing preview for an important page where the platform supports it.
- Preview, publish, retest and record substantive changes.
Use one clear source of responsibility for each setting. Installing a second SEO plugin can create conflicting output rather than improve it. If a theme or server controls the element, ask the responsible developer to inspect that layer.
Use Search Console URL Inspection when investigating how Google processed a page. Remember that stored indexing information and a live test describe different states. A plugin screen alone does not prove what Google receives.
Use the technical SEO guide for detailed diagnosis
Avoid unnecessary URL changes during a tidy-up
Changing a menu label or adding a useful link may solve the problem without moving the page. Preserve established URLs unless there is a clear reason to change them.
If a move is needed, map the old URL to an appropriate replacement, implement the correct redirect and update internal links, sitemaps and relevant metadata. Retain the previous configuration and a rollback route. Test the starting and final URLs.
Do not use a canonical tag as a substitute for every move, or redirect unrelated retired pages to the homepage. Larger changes to taxonomy, permalinks or language versions deserve a migration plan.
What this means for AI search
Google says established SEO practices remain relevant to its AI features. Clear pages and useful links support discovery and understanding, but there is no special metadata field, mandatory chunk size or schema that guarantees an AI citation.
Do not fill descriptions, alt text or hidden markup with instructions asking AI systems to recommend the business. Keep factual information consistent and apply indexing or preview controls for a defined reason. Other providers can interpret controls differently.
Google’s guidance on generative AI search features
Check whether the changes helped
First verify the intended change: the correct title is output, an accidental instruction is removed, or a customer can reach the needed page. Then review search and business data over consistent complete periods.
Compare relevant pages and queries rather than assuming that a changed description caused every movement in click-through rate. Search position, query mix, device and presentation can also change. Google may take time to recrawl and process edits.
Use completed enquiries or other useful outcomes alongside visibility and visits. Keep AI impressions, citations and referral sessions separate; Google AI impressions are already within overall Search totals. A better-looking snippet alone is not proof of more sales.
How to measure SEO success when search uses AI
A short maintenance checklist
- Does each important page have a clear task and an accurate title and description?
- Are robots and canonical settings intentional and consistent?
- Can visitors find important pages through relevant links?
- Do menus, categories and breadcrumbs help real browsing tasks?
- Are sharing details accurate without unsupported promises?
- Have URL changes been planned and tested rather than made cosmetically?
Sources and further reading
Guidance reviewed on 16 October 2026. Check original documentation before implementing indexing or presentation restrictions.


