There is a failure mode that hides better than almost any other technical SEO problem, because nothing about it looks wrong. Every menu works. Every footer link works. Click anything and you go where you expect. And yet a crawl shows large parts of the site unreachable, internal authority pooling uselessly on the homepage, and important pages either undiscovered or crawled as though they were orphans. The cause is that the links are not links. They are elements with JavaScript attached that move the browser when a human clicks them, and to a crawler they are invisible.

I found exactly this on a large marketplace audit: anchor elements throughout the navigation and footer that carried no href attribute at all, relying entirely on scripted navigation. For users, flawless. For search engines, an entire layer of the site’s internal link graph simply did not exist, which meant no discovery and, just as importantly, no authority flowing to the pages those links pointed at.

Why This Is Not the Same as the JavaScript Rendering Problem

These two issues get conflated constantly, and separating them matters because the diagnosis and fix are different. Client-side rendering is about content: the page’s text and products are drawn by JavaScript after load, so a crawler may receive an empty shell. The problem is what is on the page. This issue is about connections: the page content may be perfectly rendered server-side, but the links between pages are not real links, so the crawler cannot travel. The problem is how pages join to each other.

A site can have either problem independently. A perfectly server-rendered site can have unfollowable navigation, and a site with excellent crawlable links can serve empty JavaScript shells. Sites with both are common, and the symptoms overlap enough that teams often fix one, see partial improvement, and never realise the other is still capping them.

What Actually Breaks

How to Detect It in Ten Minutes

The good news is that this is trivially checkable once you know to look, and the checks require nothing more than a browser and, ideally, a crawler.

The middle-click test is the one I recommend to non-technical owners, because it takes five seconds and needs no tools. If middle-clicking your main menu items does not open them in new tabs, you probably have this problem, and it is worth escalating to whoever built the site before another quarter of SEO investment is spent on a site that cannot distribute it.

The Fix, and How to Talk About It With Developers

The fix is conceptually simple and usually straightforward to implement: every navigational link should be a genuine anchor element with a real href pointing at the destination URL. JavaScript can still do whatever it needs to do, intercept the click, animate a transition, load content dynamically, but it should enhance a working link rather than replace it. This is a well-established principle in front-end development, and most modern frameworks support it directly through their routing components, which output real hrefs by default when used correctly.

What to Expect After Fixing It

The recovery pattern here is one of the more dramatic in technical SEO, because you are not improving a signal, you are switching one on. Crawlers begin discovering pages through navigation, which usually shows up first as a jump in crawled and indexed pages in Search Console. Internal authority begins flowing to pages that had been receiving none, and the pages that benefit most are exactly the ones your navigation was always supposed to be supporting, category and hub pages, which often move over the following weeks and months.

The effect compounds with everything else. Internal linking strategies that were void suddenly take effect. Crawl depth improves across the board without touching the architecture, because paths that were broken now function. Pages that seemed stubbornly stuck sometimes move without any change to the pages themselves. That is the signature of an unblocked structural problem: broad improvement in places you did not directly touch.

The wider lesson is worth carrying into every technical audit: check that the mechanisms you are relying on actually exist before optimising them. An internal linking plan assumes links are links. A crawl budget strategy assumes crawlers can travel. When the foundation is quietly absent, months of sophisticated work sit on top of nothing, and the reports will not tell you, because the reports measure the plan rather than the implementation. Verify the plumbing first. It is rarely the interesting part of the job, and it is regularly the part that decides the outcome.

Where This Problem Comes From, and Why It Keeps Returning

Understanding the origin helps you prevent recurrence, because this is rarely incompetence. It usually starts with a legitimate front-end need: a mega menu that must open and close smoothly, a mobile drawer, a filter panel that updates without a page reload, an animated page transition. A developer solves the interaction with a click handler, the interaction works beautifully, and nobody notices that the underlying anchor lost its href along the way, because nothing visible broke. It is a side effect, not a decision, which is exactly why it survives code review.

It returns for the same reason: the next redesign, the next framework migration, the next well-meaning refactor of the navigation component. Unless the requirement is written down as an acceptance criterion and ideally checked automatically, it is one refactor away from coming back. Teams that have been burned once usually add a simple automated test that asserts every navigational anchor has a non-empty, non-hash href, which takes an afternoon to write and prevents years of silent damage.

The Wider Family of Links Crawlers Cannot Follow

The common thread across all of these is a single question worth adding to every technical review: can a crawler reach this page by following markup that exists in the served HTML. If the honest answer is no, then whatever else is true about that page, it is operating at a disadvantage no amount of content quality will overcome.