The Backlink Black Hole: Why 90% of Links Never Get Indexed

The pitch arrives with a name attached: the backlink black hole. Ninety per cent of your links, it says, have fallen into a void Google will never look into, and every week you wait is money already spent and quietly wasted.

It is effective marketing because it is built around something real. A backlink not indexed is a backlink not counted, and most people who buy links have never checked which of theirs are which. But the number is doing work the evidence cannot support, and so is the word “never”. Both are worth taking apart before you act on the fear.

Nobody has counted the links

Start with the obvious question: where does 90% come from? Trace it back and it dissolves. There is no public study that samples backlinks at scale, resolves each linking page, and reports what share of them sit in Google’s index. Not one. The figure circulates because it is repeated, not because it was measured.

The nearest real number belongs to a different question. Ahrefs’ search traffic study examined a sample of over a billion pages and found that 90.63% of them received no organic search traffic from Google. It is almost certainly the ancestor of the backlink claim. But read what it measures: traffic, on pages in general. Not indexation, and not the pages that carry your links. A page can be perfectly well indexed and still get no visitors, because ranking eleventh for an obscure query produces no clicks.

So the honest position is that the 90% figure is not wrong about backlinks. It is silent about them. Someone borrowed a traffic statistic, changed the noun, and sold it back to you as a crisis.

Why the story still feels true

Dismissing the number is not the same as dismissing the worry, and the worry is well earned. Three things make the gap real, and a fourth makes it feel bigger than it is.

The pages carrying your links are rarely a site’s best pages. A guest post lands in an archive, a mention in a roundup, a profile in a directory: the shallow, recently created, weakly linked parts of somebody else’s site, which is precisely the material Google is slowest to store. Google’s description of how Search works is explicit that crawling a page does not commit it to being indexed.

You also do not own those pages. Search Console answers for properties you control, and a linking site is not one of them, so the ordinary tool is unavailable exactly where the question is hardest.

And nothing notifies you when the situation changes. A page indexed in week six looks identical, from your side, to a page that will never be indexed at all, because you were not watching either one.

That is the fourth thing, and it is the engine of the whole story: the unknown fraction of a portfolio gets rounded up to the worst case, because an unmeasured link and a dead link feel the same. The void is partly an observation problem wearing the costume of a catastrophe.

A backlink not indexed is not a backlink lost

The framing’s real sleight of hand is the word “never”. It compresses three quite different states into one permanent loss.

Not yet. The linking page is in the queue. Google has either not crawled it or has crawled it and not yet decided. This is the most common state for a new placement, and it can persist for weeks. Google’s guidance on asking it to recrawl a URL notes that repeating the request does not move it up the queue.

Not ever. The page carries a noindex rule, sits behind a robots block, or is thin enough that Google has already declined it. No amount of submission changes this, and the useful response is not persistence but a conversation with whoever sold you the placement.

Not any more. The page was indexed and later dropped out: the host restructured, pruned, or was re-evaluated. Your link counted for a while and then stopped, which is invisible to anybody who checked once at the start and never again.

These three need different responses, and the black hole pitch prevents you from distinguishing them, because undifferentiated dread converts better than a diagnosis. The same instinct produces the guaranteed indexing rates on sales pages in this category. Sit with such a promise: it is a percentage attached to a decision Google makes, offered by a company with no relationship to Google. Delivery can be promised. The verdict cannot.

Measure your own portfolio instead

The number that matters is not the industry’s but yours, and unlike the industry’s it can actually be obtained.

Export your linking URLs, all of them, and check each against live Google results. What comes back is an indexation rate for your link building specifically, the arithmetic covered in what indexation rate measures applied to pages you do not own. Segment it by vendor, by placement type, by month, and the aggregate stops being a mystery and becomes a purchasing decision. Directories lagging well behind editorial placements is an actionable fact. “Most links never count” is not.

Two details make the exercise worth doing properly. The first is dates: an “indexed” verdict with no date tells you little three months later, whereas “indexed on 14 October” tells you when a placement began earning. The second is that submission and indexation are separate events, and a tool that flips a URL to green on delivery is reporting the first while you are watching for the second. We laid that out in submitted is not indexed.

This is the job AIndexify was shaped around. You paste linking URLs or upload the list as a CSV, up to 10,000 at a time, and each is checked against live Google results on a schedule, with submitted and indexed reported as two separate events. No Search Console connection is needed, which is what makes it usable for pages you did not build. A URL carrying noindex, returning a 404, or blocked in robots.txt is caught before any check runs and comes back as “Couldn’t check” with the credit refunded. That is a diagnosis in itself: the placement was never going to count, and you can say so to whoever sold it.

What the honest version of the claim looks like

Some proportion of your backlinks is not in Google’s index today. That proportion is unknown until you look, it varies by where the links came from, and part of it will resolve on its own. None of that is a black hole. It is an ordinary gap between what you bought and what has been recognised, and it closes for some links and not others.

For the mechanics behind the “not yet” and “not ever” cases, the guide to Google indexing covers what makes a page acceptable to the index and what keeps it out.

The one thing worth refusing is a statistic offered as a reason to hurry: your link list is a finite set of URLs, each with a checkable status. See where your links actually stand: submit your URLs and get the answer with dates attached.