Skip to main content
WCAGrules
Quick navigation

Multiple Ways

Any page in a set of pages has to be findable more than one way, and W3C documents six routes. Links to related pages, a table of contents, a site map, a search function, a list of links to every other page, and a home page that links them all. Two of the six is the bar, and most sites already have the first one without trying, so the usual failure is being one route short rather than having none. One exception is written in, and it is the one a shop needs. A page that is the result of a process, or a step in one, sits outside the rule entirely. Your checkout steps, your order confirmation and your search results page all live there, and building a second route to a confirmation page would be strange.

Why it matters

A single path is a bet that every visitor's head works the way your information architect's does, and that bet loses all day. Somebody with a cognitive disability can find a deep menu impossible to hold in mind while search feels easy. A screen reader user often prefers search outright, because it beats working down a menu tree link by link. Somebody using magnification is looking at a slice of a large navigation bar at a time and would far rather type. Nobody is locked out by having one route, which is the honest version of this rule. They are slowed, and some of them give up, and giving up looks exactly like being locked out in your analytics.

Who this rule protects

This affects people with cognitive disabilities who find deep hierarchies hard to hold in mind, screen-reader users for whom search beats a menu tree, magnification users reading a large navigation bar a slice at a time, and people who would rather move through a site in sequence, page after page.

How to check it yourself

  1. Pick a deep page, such as one product or one article, and count the routes that actually reach it. The menu path, the search, a site map, a category index, and sequential links if you have them.
  2. Run the search and check the result. The rule is satisfied when the search takes the user to the page or to a list of links containing it, and a search that returns nothing for a page that exists is not a route.
  3. Rule out the exception before raising a finding. A page that is a step in a process, or its result, is outside this criterion.
  4. Check the site map against three conditions. It has to link to every section, present the site's real organization, and contain no dead links. Fail any one of the three and it has stopped being a site map.
  5. Check the way back. Every page listed in a site map should link to that site map, or the map is reachable from one place and nowhere else.

Failures we see most often

  • Content reachable only by clicking down three levels of menu, on a site with no search, no site map and no index.
  • A search box that silently skips whole sections, so pages that exist can never be found by typing their names.
  • A site map nobody has updated in two years, listing sections that were renamed and linking to pages that were deleted.
  • A blog on its own subdomain with its own navigation, treated as covered by the main site's search, when it is a separate set of pages with an obligation of its own.

Who this one is for

Read from this rule's own note above, so the grouping and the note cannot disagree.

How to fix it

  • Add working on-site search alongside the navigation you already have. A link to a search page counts as much as a form in the header, so this can be one page rather than a header redesign.
  • Make the search forgiving. Spell-checking, word stemming and synonyms are all named as accessibility improvements to search, and the spell-check is the one that matters most for dyslexic users and for anyone typing on a phone.
  • Publish a site map that lists every section, matches the real structure, and links back from the pages it lists. It also satisfies 2.4.8 Location at AAA on its own, so one artefact covers two rules.
  • On a small site, link every page from the home page and link the home page from every page. That single set of links is your navigation, your site map, and your 3.2.3 Consistent Navigation answer at the same time.
  • Think twice before listing every page on every page. It works on a small site, and it creates a repeated block that 2.4.1 then requires a way past.
Step-by-step fix guides (8)

Passes vs. fails

A general illustration of the pattern rather than a test of 2.4.5. Passes: headings exist in the code. Fails: one flat wall of text.

Passes

The same page comes back the moment somebody types 'refund' into the search box, and it is listed in the site map under Policies.

Fails

The refund policy is reachable only through Footer, then Legal, then Policies, then Refunds.

In audits and lawsuits

No automated rule exists for this criterion, and that matters, because the failure is almost never a missing search box. It is a search box that quietly excludes half the site. We check that the second route actually finds the page, not that the control exists. Scope catches people out too. The unit is a set of pages sharing a purpose and an author, not a domain, so a checkout flow is its own set, a blog on a subdomain with its own navigation is its own set, and every language version of a site is its own set carrying its own obligation. A small site has an easy answer W3C states outright. Five pages, each linking to all the others, plus next and previous links, and it is done.

Go somewhere useful

Find tools, resources and your workspace.

29 destinations