WebsiteDesignOutsource.com research

Navigation Depth and Findability in Outsourced Website Projects

A bounded research model for judging whether important pages are reachable without mistaking shallow navigation for good information architecture.

August 14, 2026 Navigation Depth and Findability in Outsourced Website Projects editorial illustration

Navigation depth is often reduced to a number of clicks from the homepage. That number can be useful for locating outliers, but it is not a quality score by itself. A page two clicks away can be harder to find than a page four clicks away if the labels are vague, the hierarchy is inconsistent, or the visitor starts from a search result. An outsourced design team needs a richer record that distinguishes path length from findability.

Define the starting point

There is no single site-wide depth. A visitor may start at the homepage, a service page, a blog article, a search result, or a bookmarked URL. For each important page, record the starting point used in the test. Then record the shortest menu path, the shortest contextual-link path, and whether the page is discoverable through the sitemap. These are different signals. A sitemap helps crawlers discover URLs; it does not guarantee a human can understand the site.

Google Search Central recommends crawlable links with discoverable destinations. The practical implication is to inspect the actual anchor text and href, not merely a visual design. A button that changes state without exposing a meaningful link may support a visual prototype while leaving the route ambiguous to a crawler or keyboard user.

Labels carry more information than depth

A label should identify the destination in language a visitor can predict. Services, industries, and resources may be legitimate groups, but a team should test whether the group name matches the question a visitor has. If a page is placed under “Solutions” while the audience searches for “website maintenance,” record both the organizational label and the user language. The gap is a content and navigation finding, not evidence that one naming system is universally correct.

Card grids create a similar problem. A grid may show many destinations at the same visual level, yet headings, order, and link text determine whether the group is usable. Test the list with the images hidden and with link text extracted. This catches designs that depend on color or photography to communicate hierarchy.

A testable review frame

For each selected task, note the scenario, start route, target route, first expected action, wrong turns, completion condition, and evidence. A reviewer should be able to repeat the task without a private explanation from the designer. Record whether the person used the main navigation, a footer link, an internal search feature, or a contextual link. Do not combine these paths when comparing results.

Accessibility changes the review frame. WCAG requires keyboard operability and a meaningful focus order. A mouse-only path can therefore report success while the route remains unusable for another user. Check focus visibility, submenu behavior, headings, skip links, and the name of each interactive destination. The HTML and ARIA guidance should be treated as implementation references, not decoration added after the route is chosen.

Evidence from search and internal links

Search queries can reveal how visitors name a need. Internal link data can reveal which pages the site chooses to recommend. Neither one proves that a page is the best answer. Use them together with page content, support questions, and task testing. When evidence disagrees, keep separate hypotheses. This avoids redesigning the entire navigation around a single noisy metric.

Interpretation and limitations

A shallow tree may hide excessive choice. A deep tree may be appropriate for a large service catalogue if each level has a clear purpose. Click counts also vary with device, viewport, authentication state, and whether a visitor already knows the destination. Report the test conditions and the route inventory date. Do not turn a small task sample into a claim about all users.

Methodology

This article synthesizes W3C, WHATWG, Google Search Central, the Sitemaps protocol, and usability research on information scent and navigation. Its test frame is a practical recommendation for outsourced website work. It does not predict search rankings or conversion rates.

Test path quality, not only path length

For an important route, create two test paths: one beginning at the homepage and one beginning at a realistic entry page such as a service article or search result. The first reveals whether the main information architecture exposes the destination. The second reflects how visitors may actually arrive. Write the task in visitor language and avoid naming the destination route. If the participant already knows the label, the test measures memory rather than findability.

Record the labels encountered and the point at which the participant hesitates. A route may be reachable through “Resources” but effectively invisible to someone looking for evidence about a service. That finding does not prove that “Resources” is wrong in every context. It shows a mismatch between a group label and a particular task. The appropriate response might be a clearer sublabel, a contextual link, a revised heading, or a different grouping. The evidence should inform that choice instead of prescribing one universal menu.

Balance navigation and context

Global navigation is useful for stable destinations, while contextual links explain why a next page matters. Build a relationship record for high-value routes: source page, destination, link wording, visitor question, and whether the destination supplies the promised answer. This is more informative than counting links or demanding that every page sit at the same depth. Pagination and archives need the same discipline: verify later pages, preserve older entries, and state the deterministic tie-breaker for same-day records. The recommendation should name the observed condition, tested task, and proposed change. It should not claim that reducing one click will increase enquiries or rankings without separate evidence.

Key Stats

Compare depth with findability evidence

For each priority route, record the homepage path, the shortest contextual path, the starting route used in a task, and the labels encountered. This separates structural reachability from human recognition. A route can be two clicks away through an internal label that a visitor does not understand, while a longer contextual path can be clearer because each link answers a question. Do not average these observations into one site-wide score.

Search entry pages deserve special attention. A person who lands on an article may never see the homepage navigation before needing a service explanation or related evidence. Add links where the next question naturally arises, and make the visible label describe the destination. Test the link from the reader’s current context rather than assuming that a global menu will repair a missing relationship.

When reorganizing a tree, preserve route identity and check all affected links. A new group name can improve recognition while an old bookmarked route remains valid. A redirect, canonical, or archive change needs its own review because discoverability and URL continuity are different concerns. Sitemap presence helps discovery by crawlers, but it does not demonstrate that a person can choose the correct destination.

The conclusion should state the task and conditions that support it. “The service page was found from the homepage by three participants at a named viewport” is bounded. “The site has excellent navigation” is not. This distinction gives an outsourced team an actionable change while protecting the owner from a false performance guarantee.

WCAG 2.2 uses 4 accessibility principles (W3C)

An XML sitemap uses a urlset document root (Sitemaps.org)