Do Internal Links Help SEO? Ten Links That Did Nothing

By Ugur Saritepe · September 11, 2026

Do internal links help SEO? Only when the pages doing the linking have search presence of their own. I audited a page of mine with ten inbound internal links and one click in 90 days: five of the ten linking pages had zero Google impressions themselves, and links from pages Google never serves passed on nothing.

The page ten links could not lift

The page is /best/best-zombie-games on game-scout.app, the games directory I run and write about openly. Over the 90 days ending 10 September 2026 it earned 1 click from 409 impressions at an average position of 28. On paper its internal linking looked healthy: it belongs to a curated topic cluster, and Search Console listed ten of my own pages linking to it. That is more inbound links than most pages on the site get. I had spent real effort making sure no page was orphaned the way nine of my lists once were, so the reach made no sense to me. This is what I typed:

one per-page pull for each of the ten linking pages, window 2026-06-13 to 2026-09-10

5 of 10 linkers (resident evil, indie horror pc, coop horror pc, the browse hub, a vs article): 0 impressions
/browse                        1 impression    0 clicks
/best/best-cyberpunk-games    14 impressions   0 clicks
/best/best-horror-games       58 impressions   0 clicks   position 21.9
/categories                   62 impressions   0 clicks
/best                         63 impressions   1 click    position 7.9

An impression is one appearance in Google's search results, so zero impressions over 90 days means Google never showed that page to anyone. Half of my ten links sit on pages in that state. The links exist in my site's HTML, Google can crawl them, and that is the entire story of what they do.

Where the ten links came from, and why nobody chose them

The uncomfortable part is that no one picked those ten linkers, including me. The cluster this page belongs to computes its related-list links: each of the 12 member pages links to the next four members in the list, cyclically, plus two cross links into neighboring clusters. Inbound count is pure index arithmetic. My zombie page has exactly four sibling links because every member has exactly four, and which four pages link it is decided by its position in an array, not by which pages have anything to pass on. Its slot happened to hand it the cluster's weakest members.

That is not an exotic setup. Every related-posts widget, every auto-generated hub, every "you might also like" block works this way: link coverage is guaranteed, link quality is accidental. My audit of the site's orphan pages had checked that every page has inbound links. It never asked whether the linking pages were worth anything in search. Being in the cluster is a fact about my site's link graph. It is not a fact about search.

What internal links actually do, and what they cannot

Internal links reliably do one job: discovery. Google's own documentation is mostly about making links crawlable so Googlebot can find your pages, and my links do that job fine. The zombie page is crawled and indexed. Discovery was never its problem.

The second job people expect, passing authority so the linked page ranks better, is the one that failed here, and my data says the failure mode is specific: a linking page Google never serves has no standing to lend. I want to be careful about what this one case proves. The zombie page also targets a competitive query and averages position 28, so better links alone might not rescue it. What the audit does show is that a page can pass every link-coverage check you have and still take all of its links from pages with nothing behind them, and no tool warns you, because link audits count edges, not the search performance of the pages the edges start from.

Check what your internal links are worth in five minutes

The test is one question asked ten times: for each page linking to your underperformer, what did that linking page itself earn in search? You need the list of linking pages (Search Console reports which of your pages link to a URL, or take them from your own template logic) and a per-page performance pull for each one:

• Linker earns impressions and clicks:a real link. Keep it, and consider linking from higher in that page's body.
• Linker earns impressions, no clicks: it has some standing with Google. The link is doing quiet work; the linker has its own problem worth a separate look.
• Linker has zero impressions over 90 days: the link is a crawl path and nothing more. Do not count it in your coverage, and do not add more links like it.

The bad output looks like mine above: most of the list in the third bucket. The good output is boring, a table where the linking pages earn hundreds of impressions each. On the same site, my strongest cluster's pages link each other while each earning thousands of impressions, which is what makes those links matter.

The fix is placement, not more links

For my zombie page the fix is not adding links, it is moving the page's slot so its computed links come from the two cluster members that actually get served, one of which earned 4 clicks from 78 impressions in the same window. Same number of links, different sources. We have not applied that reorder as I write this; the page still takes its four links from the cluster's weakest members, and I will report what happens when we do.

The honest limits of this case: one page, one site, one 90-day window, and link value measured by proxy, through the linkers' own search presence rather than a controlled experiment. What it settles for me is the audit itself: I now check where a page's links come from before I count them.

This post is part of our AI-assistant SEO series, written from the SEO work we do on our own sites. For a five-minute check of the basics on any site, the free triage report runs eight checks from a URL alone, no account needed.

FAQ

Do internal links help SEO?

Internal links help Google discover pages and understand how they relate, but only links from pages Google actually serves in search carry any weight a searcher will feel. On one of our pages, ten internal links produced no measurable lift because five of the ten linking pages had zero Google impressions themselves over 90 days. Count where your links come from, not how many you have.

Do internal links from pages with no traffic count?

They still help Google find the page, so they are not worthless, but that is where their measurable effect ended in our data. A page linked from ten others earned one click in 90 days when half of those linkers had never appeared in search results at all. If the linking page has no search presence, treat its link as a crawl path, not a boost.

How many internal links does a page need?

There is no working minimum, and our case suggests the count is the wrong measure entirely. The page in our audit had ten inbound links, more than most pages on the site, and stayed invisible. One link from a page that ranks and gets served beats five links from pages that do not, so audit the linkers' own search performance before adding more links.

In your AI assistant

Ask your AI about your own site

TrafficTriage plugs into Claude, Cursor, and ChatGPT. Your assistant runs the 8 checks, reads your Search Console data, and can even check your site every morning. Free account, no API key, nothing to install.

Sign up freeSee how it works