How to Structure a Website So Search Engines Can Read It

You can write excellent content on a badly structured site and watch it go nowhere. Structure is what tells a search engine which pages matter, how they relate, and which one to show for a given query. It is also the part that is genuinely painful to change after launch, so it is worth thirty minutes now.

Keep the hierarchy shallow

Every important page should be reachable from the homepage in three clicks or fewer. Deep nesting buries pages, dilutes their authority, and makes them harder for anyone — human or crawler — to find. If your structure will not flatten, your navigation is compensating for it, which is the problem described in navigation design.

Decide URLs once

• Short, lowercase, hyphen-separated, no dates, no IDs, no stop words.
• One canonical pattern per content type: /services/web-design, /blog/post-title.
• Pick www or non-www and redirect the other with a 301, permanently.
• Same for trailing slashes — choose, enforce, never mix.

Every URL you change later costs a redirect and some ranking. Getting the pattern right the first time is free.

One H1, then a real heading hierarchy

One <h1> per page, describing what the page is. Sections use <h2>, subsections <h3>, and you never skip a level for styling reasons. Headings are how both crawlers and screen readers build a mental model of the page — the same overlap noted in why accessibility is not optional.

Group content into topic clusters

One substantial pillar page per major topic, with narrower posts around it linking back. Search engines infer expertise from the density and coherence of those connections, not from any single page. This post is part of such a cluster: it links to navigation, architecture, and multilingual SEO because they genuinely belong together.

Internal links are the cheapest SEO you own

They spread authority, they define relationships, and they keep people on the site. Rules that work: link with descriptive anchor text rather than "click here", link from strong pages to the ones you want to rank, make sure every page has at least one inbound internal link, and prune links that exist only to hit a quota.

Get the technical basics in place

• XML sitemap listing every page you want indexed, and nothing you do not.
• robots.txt that blocks admin and API paths but never your CSS or JS.
• Canonical tags on every page, absolute, self-referencing by default.
• Structured data — Organisation, Article, Breadcrumb, and Product where relevant.
• Server-rendered HTML, so content exists without JavaScript executing first.

Handle duplication before it happens

Filters, sort parameters, pagination, and tag archives quietly generate near-identical pages. Decide up front which ones get indexed, canonicalise the rest, and keep parameter combinations out of your sitemap. For multi-language sites, add hreflang properly — the details are in the multilingual SEO guide.

Speed is part of structure

Crawl budget and rankings both respond to performance, and a site that is slow to render is slow to index. Which is why Core Web Vitals and the usual performance mistakes belong in the same conversation as your URL plan.

Planning a rebuild? Structure is the cheapest thing to get right and the most expensive to fix. Talk it through with me first.