The 3-click rule in SEO: myth, useful heuristic, or both?

The rule entered circulation through early-2000s web design writing as a usability heuristic: users get frustrated and leave if they cannot find something within three clicks.

In 2003 Joshua Porter tested it directly, giving users tasks on sites where the answer required varying numbers of clicks. Success rates did not fall off after three clicks, and users did not abandon at that threshold. What predicted success was whether each click felt like progress — whether the label they clicked matched what they got.

So as a usability law it is dead. What kept it alive is that SEO picked it up in the mid-2000s, where a different mechanism made it look right.

The SEO version is not really about user patience. It is about two mechanical facts that do favour shallow pages.

1. Authority decays per hop. Each link divides a page's authority among its targets, so a page three hops from a well-linked homepage receives a small fraction of what a page one hop away receives — the mechanism in PageRank.
2. Crawlers reach shallow pages first and more often, so deep pages go stale sooner when the crawl allocation runs out.

Both are real, and both are gradients rather than thresholds. Nothing changes at exactly three. The number is a rounded-off intuition about a smooth curve, which is why it feels approximately right and fails precisely.

On a small site the rule is harmless, because everything fits within three clicks anyway. On a large site, enforcing it causes a specific and well-understood failure.

To reach 100,000 pages within three clicks you need roughly 46 links on every page at every level, and in practice teams achieve it by bolting an enormous menu or footer onto every page. The consequences:

• chevron_rightAuthority stops differentiating. When every page links to hundreds of others, each link carries very little and internal authority becomes nearly uniform — the site can no longer signal which pages matter.
• chevron_rightDepth metrics become meaningless. A shallow average produced by a mega-menu describes the template, not the architecture.
• chevron_rightUsers get worse navigation, not better. A 300-item menu is harder to use than a clear two-step path, which is what the original research actually found.

Replace the universal rule with two targets that scale, and one question about labels.

1. A hard ceiling for pages that earn revenue or represent priorities: 2 to 3 clicks. This is where the 3 was always pointing, and applied to this subset it is a good number.
2. A soft ceiling for the tail, set by site size. On a 100,000-page site, depth 5 for an old blog post is entirely normal — see ideal click depth for the arithmetic.
3. For each hop on the path to an important page, ask whether the link label predicts what the user gets. That is the finding that replaced the rule, and it is still the best navigation test available.

  3-click rule            depth target
  ---------------------   ---------------------------------
  applies to every page   applies to pages that matter
  fixed at 3              scales with site size
  met by adding menus     met by contextual links + hubs
  measured as an average  measured as "deepest valuable page"

There is a reading of the rule that survives all of this: if something important on your site takes more than three clicks to reach, that is a prompt to go and look at why.

Often the answer is benign — a large catalogue with sensible hierarchy. Sometimes it reveals a page nobody links to except through a menu, a category that was never given a hub, or content that was published and forgotten. Used as a trigger for investigation rather than a threshold to enforce, three clicks is a perfectly good number.

And the fix, when there is one, is almost never a menu entry. Adding contextual links from relevant shallow pages reduces depth for the pages that deserve it without flattening everything equally — the approach in how to improve crawl depth.