Website Structure: How to Plan Pages That Build Trust

Does your site walk visitors to your best work, or leave them to find it?

Let's map your structure before we talk about visuals.

Is website structure a menu, a sitemap, or something else?

Website structure is the way your pages are grouped, ranked and linked to each other. It answers two questions for every page: where does it belong, and where can a visitor go from here?

Most teams picture it as a sitemap diagram. A box for the homepage, a row of boxes underneath, smaller boxes below those. That diagram is useful, but it only shows one layer of the real thing.

It is not your menu

Your main menu is a shortcut list. It surfaces a handful of destinations you consider important. A forty-page site might show five of them. The rest of the structure lives in footers, in-page links, related content and calls to action.

It is not your XML sitemap

The XML sitemap is a file for crawlers. It lists URLs you want found. It says nothing about which page should come first, or which one should convince a skeptical buyer.

It is a set of decisions about priority

Every structure is a ranking, whether you meant to rank or not. The pages one click from the homepage are the ones you are telling people matter. The pages buried four levels down are the ones you are telling them to skip.

That is why structure belongs in the strategy phase. It is a positioning tool disguised as a technical task.

Your site has three maps, not one

This is the idea most guides skip, and it changes how you plan. A site is organized by three overlapping maps. They share pages, but they answer different questions.

The address map

This is the URL hierarchy: /en/projects/ for case studies, /en/blog/ for articles, and so on. Its job is stability. A good address still makes sense in five years and survives a full redesign.

Google's guidance on URL structure asks for readable words over ID numbers, and hyphens between words rather than underscores. It also warns that URLs are case sensitive. Simple rules, but they save painful migrations later.

The navigation map

This is everything a visitor can click to move around: menus, footer, breadcrumbs, cards, contextual links. Unlike the address map, it can and should evolve. You can add a shortcut to a new case study tomorrow without moving any URL.

Google itself separates these two maps. Its breadcrumb documentation recommends trails that reflect a typical user path rather than mirroring the URL structure. The address and the journey are allowed to differ.

The proof path

This is the map almost nobody draws. It is the order in which a visitor meets your arguments: the promise, then evidence, then reassurance, then a clear next step.

A site can have clean URLs and a tidy menu and still fail here. The proof sits in the wrong place, or the next step never comes. When we review brand sites, this is the map that is usually missing.

Why the third map matters more for deep pages

Breadcrumbs help a visitor understand where they are. They are still rare. According to the HTTP Archive (Web Almanac 2025, SEO chapter, 2025), breadcrumb structured data appears on only 28% of desktop inner pages. Most deep pages therefore rely on the content itself to explain where the visitor stands.

The Nielsen Norman Group adds a useful nuance in its breadcrumb guidelines. Breadcrumbs should reflect the site hierarchy, not the visitor's history. They are also unnecessary on flat sites only one or two levels deep. On a small brand site, the proof path has to do that orienting work instead.

Flat or deep: which hierarchy fits a premium brand site?

For most brand sites under fifty pages, a mostly flat structure wins: few levels, clear groups, strong links between them. Depth becomes useful only when categories stop being distinct for the people browsing them.

The trade-off in plain terms

The Nielsen Norman Group framed the choice clearly in its article on flat versus deep hierarchies. Flat structures put content within reach but can overwhelm with long menus. Deep structures keep each menu short but bury specific content under generic labels.

Its 2013 examples still make the point. One hospital site listed 32 medical services directly in a single menu. Another placed disease-specific pages three levels below the homepage. Neither is wrong in itself. The question is whether visitors recognize the categories at a glance.

Where long menus break down

Long lists are where flat structures fail. According to Baymard Institute (Homepage and Category Navigation UX 2026, 2026), 51% of leading ecommerce sites do not divide categories into manageable chunks. In its testing, most participants felt overwhelmed past roughly ten subcategory options.

A brand site rarely has that many categories. But the same logic applies to a portfolio filter with twenty tags, or a services menu listing every deliverable. Group first, then show.

What a premium site usually needs

In practice, a premium brand site tends to settle into three groups of pages.

• Offer pages, which explain what you do and for whom.
• Proof pages, such as case studies, client work, opinions and awards.
• Conversion pages, from a contact form to a booking link.

Everything else hangs from these three. A blog feeds proof. An about page feeds trust. A careers page sits outside the sales path entirely and should not compete with it in the main menu.

How does website structure affect SEO?

Structure affects SEO mainly through internal links and clarity, not crawl capacity. For a site of a few dozen pages, the question is whether Google understands what each page is for, not whether it can reach it.

Crawl budget is not your problem

Many teams worry about crawling. They usually should not. According to Google Search Central (Optimize your crawl budget, updated 2026), that guide mainly targets sites with over a million pages that change weekly, or over 10,000 changing daily. Google calls these rough estimates. A thirty-page studio or firm site is nowhere near either threshold.

Internal links are your real signal

Links are how both visitors and crawlers move. According to the HTTP Archive (Web Almanac 2025, SEO chapter, 2025), the median desktop page holds 43 internal links, and the median mobile page holds 39. Every page is already a hub, whether you planned it or not.

The question is what those links say. Forty links to legal pages and social profiles send a weak signal. Six well-chosen links from a case study to related work, a matching offer and a contact page send a strong one.

Sitelinks reward a readable structure

Google's sitelinks documentation explains that those extra links under a result are generated automatically from your site's link structure. Google lists a few things that help: clear titles, easy navigation, links to important pages from relevant pages, and short, relevant anchor text.

You cannot order sitelinks. You can only make your structure easy to read. For a brand searched by name, that is often the difference between one result and a block of shortcuts to your best pages.

Labels matter as much as links

A link labeled "Solutions" tells nobody anything. A link labeled "Websites for architecture firms" tells visitors and search engines exactly what waits on the other side. Specific labels are a usability fix and an SEO fix at the same time.

Not sure which pages your site is actually pushing forward?

Let's look at your structure together.

A five-step workshop for planning your site structure

This is the sequence we use before a single screen is designed. It works for a new site and for a redesign. Each step produces something you can review and sign off.

Step 1: Collect the questions buyers actually ask

Start with sales calls, inbound emails and search queries, not with your org chart. List the questions a prospect asks before they hire you. What do you do? For whom? Have you done it before? What does it cost? How does it work?

Each recurring question needs a clear home. If no page answers it, that is a gap. If three pages half-answer it, that is a structure problem.

Step 2: Sort every page by job

Put every existing or planned page in a spreadsheet. Next to each one, write its single main job: explain the offer, prove a claim, reassure, convert, or attract search traffic.

Pages with no clear job are candidates for merging or removal. Pages with two competing jobs usually need splitting. This spreadsheet also makes a strong starting point for your website project brief.

Step 3: Choose the depth, group by group

Decide how many levels each group needs. Offer pages are often a single level. Proof pages may need a list page plus individual case studies. A blog may need a few topic groups.

Run a quick card sort if you are unsure. Hand a few clients or colleagues the page titles and ask them to group them. Their groupings will rarely match yours exactly, and the differences are the useful part.

Step 4: Write labels and addresses together

Name each group in the words your buyers use. Then write its URL in the same plain language, short and lowercase, with hyphens between words. Labels can change later. Addresses should not.

This is also where you decide what goes in the main menu. Keep it short. Everything else is reachable through pages, not through a crowded header.

Step 5: Test the bare tree, then watch the live site

Before design, run a tree test. Visitors see only your page labels in a clickable outline and try to find specific things. The Nielsen Norman Group recommends this step before layouts and content exist, since editing a spreadsheet is far cheaper than editing a built site.

After launch, keep watching. A UX audit a few months in shows where visitors hesitate, loop or leave. Structure is never finished. It should follow the business as it grows.

Structure mistakes we keep seeing on beautiful sites

These are not beginner errors. They show up on polished, award-worthy sites built by experienced teams.

Mistake 1: The case study dead end

A visitor reads a great case study, scrolls to the bottom, and finds nothing but a footer. No related project, no matching offer, no invitation to talk.

This pattern comes up constantly in portfolio reviews. The team assumes visitors will return to the menu. In reality, many will simply close the tab once the story ends. Every proof page needs a next step built into its last screen.

Mistake 2: Categories that should have been filters

Teams love neat categories. Visitors often need to combine criteria instead. According to Baymard Institute (Homepage and Category Navigation UX 2026, 2026), 79% of leading ecommerce sites use categories where filters should be used.

The same trap appears on project pages. A firm splits its work into rigid sections by sector, then a prospect wants every residential project of a certain size. A filterable list with clear tags serves that visitor far better than five separate pages.

Mistake 3: An experience that hides the map

Immersive menus, full-screen transitions and scroll-driven pages can be wonderful. They become a problem when they replace the structure instead of dressing it.

Picture a launch week. The site opens with a cinematic intro, and pages slide in through custom transitions. Everyone celebrates. Weeks later, Search Console lists case studies that Google never discovered. Google's link guidelines explain why: it reliably follows only standard anchor elements with an href. Keep the cinema, and put plain links behind it.

What our lead developer thinks

What a repositioning looks like in structure

Structure is not neutral. It tells visitors what kind of company you are before they read a headline. A repositioning that changes the logo but keeps the old structure often sends a mixed message.

The Apax project

The Apax project shows this clearly. Apax is a Paris architecture agency founded in 1982 and taken over in 2025 by Antoine Morizot. The studio Graphéine redesigned its identity and name to set a new direction.

According to the project page, the website had one goal. It had to move Apax from a traditional image to a contemporary, confident and selective practice. The result, as the page describes it, positions the agency to speak to a demanding international clientele.

Selection as a structural choice

The visuals on the project page show a homepage, an agency page and project pages. They also show three ways to browse the work: a list, a grid and a selection. That last view matters for a firm aiming to look selective. A curated entry point says "these are the projects that define us" before the full archive appears.

That is structure doing positioning work. The same projects, offered through a curated door, read differently from a long chronological list.

How our own site applies the same logic

Our own site follows the three-group logic. Its main menu holds five items: home, projects, about, merch and contact. There is no services page and no pricing page.

Instead, dedicated entry pages speak to specific audiences, such as startups, production companies and luxury brands. Each one speaks to its audience directly, so a visitor arriving from search does not need the homepage to understand the offer.

Your most common questions about website structure

Structure is the design nobody screenshots

Nobody shares a sitemap on social media. Nobody wins an award for a well-placed related-project link. Yet structure decides whether the rest of the design gets seen.

The pattern is consistent. Plan three maps, not one. Keep addresses stable and let navigation evolve. Group pages by the job they do for a buyer. Name things the way buyers talk. Test the bare tree before anyone designs a screen.

Get that right, and every visual decision afterward has a clear role. Browse our recent projects with one question in mind: where would you enter, and what would you need to see next?

Want a site whose structure sells as well as its design looks?

Tell us about your project.