Craig Campbell SEO Training & Consultancy

Tutorials

Site Architecture and URL Structure for SEO

By Craig Campbell · 18 Nov 2026

11.9 min read

Site Architecture and URL Structure for SEO

Site architecture is how you organise pages and connect them so people and search engines can understand the website. URL structure is part of that job, but it is not the whole job. A beautifully named address does little good if the page is buried, orphaned or surrounded by confusing navigation.

I start with what the business offers and what visitors need to do, then build a structure that supports both. This guide explains how I map a hierarchy, choose practical URLs and improve an existing site without turning a tidy-up into an unnecessary migration. Every example website and path below is illustrative, not a client case study.

Quick answer: what makes good site architecture?

Good site architecture groups related content logically, gives important pages clear internal routes and uses stable, understandable URLs. Visitors should be able to predict where something belongs. Search engines should be able to discover your key pages through ordinary crawlable links rather than depending entirely on search forms or a sitemap.

For URLs, I normally favour descriptive words, consistent formatting and as little unnecessary complexity as the platform allows. That does not mean every URL must be short or every page must sit directly under the root. The right structure supports meaning and maintenance without creating needless variants.

If an existing URL works, do not change it purely because another format looks prettier. Improve navigation, page relationships and internal links first. URL changes introduce redirects, testing work and potential disruption, so they need a clearer justification than a personal preference for neat folders.

Map user needs before drawing the hierarchy

Start by listing the main tasks visitors come to complete. On a shop, that could mean browsing a product type, comparing options, checking compatibility or finding delivery information. On a service website, it might mean understanding a service, assessing suitability and making an enquiry.

Group those tasks by topic and intent. A category page helps someone explore a range; a product page helps someone evaluate a specific item; a guide answers a question that may sit earlier in the decision. They can support each other without all competing for exactly the same purpose.

I combine this exercise with a content inventory. Record each important URL, its template, purpose, parent topic, indexability and current internal-link routes. Add available performance data, but do not let traffic alone define value. A low-traffic page might answer an important customer question or support a narrow commercial need.

Look for duplication in purpose, not just repeated wording. Two service pages may target the same need with slightly different names, while two similarly named guides might answer genuinely different questions. Make consolidation decisions from the content and intended journey rather than a spreadsheet rule that treats matching keywords as proof.

Finally, involve whoever maintains the website. An elegant diagram is useless if the CMS cannot support it without fragile workarounds. Agree which relationships are editorial, which are generated automatically and who decides where new pages belong.

Step by step: design a workable hierarchy

Step 1: define the main sections. Use language your audience recognises. Your internal departmental names may make sense in a board meeting but confuse potential customers. Keep the top level focused enough that visitors can scan it without interpreting a catalogue of every page on the website.

Step 2: create useful parent pages. A parent should help someone choose a direction, understand a topic or explore a range. Do not create empty category shells solely to place another keyword in a path. If the parent adds nothing, reconsider whether that layer belongs in the hierarchy.

Step 3: place pages under the most sensible topic. A page can be relevant to more than one section without needing multiple indexable copies. Choose one stable preferred URL and link to it from other relevant places. You are designing routes to content, not manufacturing a separate page for every route.

Step 4: check important journeys. Walk from the homepage to key categories, services and products. Then start from a guide or a search landing page instead. People rarely enter only through the homepage, so each page should make its local context and sensible next steps clear.

Step 5: test the labels. Ask someone unfamiliar with the internal plan where they would look for a particular item. You do not need an elaborate research programme to expose obvious ambiguity. If several people put the same item somewhere else, investigate the labelling rather than blaming them.

Step 6: plan for growth. Consider what happens when you add new products, topics or services. A structure tied to a short-lived campaign or today’s organisational chart may require repeated rebuilding. Stable audience-facing categories usually age better than navigation built around temporary internal priorities.

Choose practical URL structure rules

A URL should identify a resource reliably. I normally use lower-case words separated by hyphens, avoid unnecessary parameters for primary landing pages and keep naming conventions consistent. These are practical defaults, not a reason to rewrite every successful address on an established website.

DecisionPractical defaultReason
Word separatorsReadable hyphenated wordsEasy to understand and maintain
Letter caseConsistent lower caseReduces accidental variants
Category foldersUse where stable and helpfulExpresses organisation without forcing duplication
DatesInclude only when useful to the content modelAvoids unnecessary age cues and future moves
Tracking parametersKeep out of permanent internal destinationsLimits duplicate URL signals
Trailing slashChoose and enforce one approachPrevents competing technical versions

For an illustrative service business, /services/roof-repairs/ communicates a relationship clearly. A store might choose /products/blue-widget/ as a stable product address while linking to that product from several categories. Neither pattern is automatically right for every platform.

Be careful about embedding a changeable category into every product URL. If products move between categories, that design may create unnecessary URL changes or several addresses for one item. A stable product path can make maintenance easier while categories and breadcrumbs express the browsing relationships separately.

Do not stuff a URL with every variation of a keyword. Repetition makes addresses harder to read and rarely solves a content relevance problem. Choose a concise description of the page, then let the title, headings, main content and links explain the detail.

Also decide what happens to case variants, duplicate slashes and unwanted hostname or protocol versions. Consistent server behaviour prevents technical variants from undermining an otherwise clear structure. Document the rules so future developers do not introduce a competing convention in a new section.

Connect the hierarchy with useful internal links

A folder structure does not create links by itself. Search engines still need actual routes between pages. Use ordinary links with valid href attributes for important destinations, even if JavaScript adds interactive behaviour. A button that only runs a search or changes interface state is not automatically an equivalent discovery route.

Main navigation should surface your principal sections. Category pages should link to relevant subcategories or items. Detail pages should give users a way back to useful parents and onward to related content. Breadcrumbs can make that context clearer, especially when people arrive directly from search.

Contextual links connect pages where the relationship matters in the content. A buying guide can point to a relevant category, and a service page can point to a guide that answers a common concern. Use anchor text that explains the destination rather than repeating vague instructions such as “click here”.

I do not use a universal rule that every page must be within a fixed number of clicks. Instead, I ask whether the most important pages are easy to reach and whether depth reflects sensible organisation. A large archive can be deeper than a core service without automatically being broken.

For implementation details, see my internal linking guide. Avoid treating links as a mechanical allocation exercise. The best architecture creates useful routes that a visitor might genuinely follow, while keeping your priority pages visible and connected.

Manage filters, pagination and overlapping categories

Large catalogues create more potential URLs than most teams initially expect. Colour, size, price, brand and sort order can combine into a huge number of variants. Decide which combinations deserve searchable landing pages and which are simply interface states that help visitors narrow a selection.

Do not make every possible filter combination indexable by default. Assess whether a page has a distinct purpose, enough relevant inventory and a useful experience. Equally, do not block all filtered pages automatically if some answer clear, valuable needs better than a broad category.

Keep useful paginated listings crawlable through links. Infinite scrolling can work for users, but the underlying implementation still needs a discoverable way to access the items. Test what happens when a crawler does not interact with the interface like a human repeatedly scrolling the screen.

Overlapping categories are not inherently bad. A product can belong in several browsing groups, just as a guide can support several topics. The aim is to avoid needless duplicate pages and confusing relationships. One preferred product address can be linked from multiple useful categories without duplicating the product itself.

Think carefully before deleting low-inventory categories. A temporary shortage is different from a permanently empty section with no future purpose. Your architecture should cope with stock changes without constant URL churn. Agree the lifecycle rules with the people managing inventory and content.

Audit an existing website before changing URLs

Crawl the live website and compare it with your intended hierarchy. Look for orphaned pages, dead-end templates, unexpectedly deep important pages and navigation links pointing through redirects. Compare that crawl with sitemaps and publishing records so you do not mistake “not found by this crawl” for “does not exist”.

Inspect a representative sample manually. Crawlers can report link depth, but they cannot fully explain whether a navigation label is misleading or a parent category is useful. Follow the journeys yourself on mobile as well as desktop. Menus often behave differently across screen sizes.

Prioritise improvements that preserve working URLs. Better labels, stronger category pages, clearer breadcrumbs and contextual links may solve the actual problem without moving anything. A mismatch between the visual hierarchy and the URL path is not automatically an SEO emergency.

When changes genuinely require new addresses, treat them as a migration. Build a mapping from each old URL to its most relevant replacement, update internal links and test the redirects. My website migration guide covers the broader planning work, including keeping a record of what changed.

Do not redirect every removed page to the homepage. Where a close replacement exists, use it. Where no replacement exists, a proper removal response may be more honest than an irrelevant destination. My redirects guide explains how to make those decisions without building unnecessary chains.

Measure whether the new structure works

Before release, record the pages and journeys you expect to improve. That might mean fewer orphaned products, more direct access to a service section or cleaner discovery of a new content hub. Define the outcome so the project does not end with a diagram and no practical validation.

After release, crawl the same representative areas again. Check destination status codes, internal-link targets, breadcrumb output, canonicals and sitemap alignment. Confirm that menus work without accidental access restrictions and that old addresses go directly to appropriate replacements where redirects were required.

Monitor search performance at page-group level rather than judging the whole project from a single day’s traffic. Seasonality, demand and other website changes can obscure the effect. Search Console can help you review impressions, clicks and indexing patterns, but those signals need context.

I also look at whether visitors can complete the intended task. If the structure makes a service easier to find but the page itself remains confusing, architecture alone has not solved the business problem. Use the findings to improve labels, content and journeys rather than repeatedly reshuffling URL folders.

Common site architecture mistakes

Flattening everything under the homepage. Fewer folders do not automatically create a better site. If every topic competes in one giant menu, visitors may have less context, not more. Logical grouping matters more than making a diagram look shallow.

Building sections around the same intent. Separate pages need a reason to exist. Creating several near-identical service categories for wording variations can produce confusing navigation and overlapping content. Consolidate where the purpose genuinely matches, but preserve pages that solve meaningfully different needs.

Creating rigid silos that prevent useful links. Topic organisation is helpful; forbidding relevant links between sections is not. A guide can naturally support more than one commercial area. Do not make users return to the homepage simply because a spreadsheet says two pages belong in different silos.

Relying on search boxes and sitemaps. Internal search is useful for visitors, but important pages still need navigable links. An XML sitemap can advertise URLs without explaining their full relationship or making the website usable. It is a supporting mechanism, not the architecture itself.

Changing URLs without an operational reason. A more fashionable slug rarely justifies a site-wide move. Count the implementation and maintenance cost, including redirects and updates to internal references. Make the change when it solves a real problem, not because an audit tool dislikes a cosmetic detail.

Site architecture planning checklist

  1. List the tasks visitors need to complete and the topics they expect.
  2. Inventory existing pages and assign each a clear purpose.
  3. Design useful parent sections with audience-friendly labels.
  4. Choose stable preferred URLs and document formatting rules.
  5. Link important pages through navigation, categories and relevant content.
  6. Check mobile journeys, breadcrumbs and crawlable links.
  7. Set deliberate rules for filters, pagination and overlapping categories.
  8. Find orphaned pages and unnecessarily difficult routes.
  9. Map and test redirects before changing existing URLs.
  10. Recrawl after release and review search and user outcomes.

Keep that checklist in the publishing workflow, not just the redesign folder. Architecture degrades when every new page is created in isolation. A short decision about its purpose, parent and links is easier than repairing years of accumulated exceptions.

Frequently asked questions

Is a flat site structure always better for SEO?

No. A shallow structure can help important pages remain accessible, but removing useful hierarchy can make a large website harder to understand. Prioritise sensible grouping and strong routes to key pages. Click depth is a diagnostic clue, not a universal score that must be minimised everywhere.

Do keywords in URLs improve rankings?

Descriptive words can help communicate a page’s subject, but a keyword-filled URL is not a substitute for useful content or good links. I use readable, relevant slugs when creating new pages. I would not normally move an established page just to squeeze another keyword into its address.

Should products include categories in their URLs?

Either approach can work. Category paths may communicate context, while category-independent product URLs can remain stable when merchandising changes. Choose the model that fits your platform and maintenance needs. Then keep one preferred product address and express additional category relationships through navigation and links.

Are breadcrumbs enough to fix poor architecture?

No. Breadcrumbs help explain context and give visitors a route to parent pages, but they cannot repair empty categories, missing internal links or unclear section labels. Use them as one component of a coherent navigation system, and ensure the visible trail reflects a relationship that actually makes sense.

When should I change an existing URL structure?

Change it when there is a substantial reason, such as a platform requirement, a business restructuring or a genuinely unmanageable duplication problem. First check whether better linking and navigation can solve the issue without moving pages. If a move is necessary, plan it and test it as a migration.

A sound structure makes the website easier to use, publish and crawl. If you need to work out which architecture problems actually deserve attention first, start with an SEO audit.

Let's work together.

Whether you need an SEO partner, a consultant, a speaker for your event or training for your team — get in touch and let's chat.

Get in touch
Get in touch