I can picture the scene, because I have been on both sides of it.

Someone searched for something, found your article, opened it. And, rarest of all, read it to the end. They arrived at the final paragraph with a head full of questions they did not have seven minutes earlier, glanced down to see what came next…

And nothing came next. Footer.

They closed the tab. They were yours for seven minutes, they trusted you, and you walked them off a cliff. Am I wrong?

Let us call it what it is: you built a wall, not a corridor

This is not really a technical problem, it is a hospitality problem. When someone comes into your office and asks a question, you do not answer and then stare at your screen until they leave. You say «oh, and while we are on it, you should know that…». That is internal linking. The «while we are on it».

And I do not mean the menu. Everyone has a menu, and readers ignore it the way they ignore the fire exit signs in a restaurant where they are having a good time. I mean the link inside the sentence, at the exact moment the question is born and not three screens later, when it has already been forgotten.

It happened on my own site, this very week

Before I tell you what to do, let me tell you what I got wrong, because that is more useful.

I have four services that all concern the same website: cookie consent, analytics, site health monitoring, and support. Four pages, written with care, with «what’s included» sections, with questions and answers, with prices. Quite proud of them.

And not one of them knew the other three existed.

Someone who came in to read about cookie consent left knowing exactly one of the things I do. If the second one would have suited them, they would have had to guess. Worse: I sell them as a bundle, at a discount, and no page explained that bundle. It existed only inside the shop, for whoever wandered that far by luck.

I fixed it with two moves. I built a page that shows all four together, with a comparison table and the bundles, and I put a link to it on every product page. The home page, which had four wide slabs stacked one under the other, became a single block with four small tiles in a row. The products section lost 23% of its height.

So the visitor now sees more in less space. That is not a coincidence, it is the rule: linking usually makes pages shorter rather than longer, because you stop repeating on every page what you have already said somewhere else.

What Google actually says, in its own words

I want to be precise here, because a great deal of folklore circulates.

Google’s documentation on links says that Google uses links as a signal when working out how relevant a page is, and to find new pages to crawl. Two jobs, not one. And it says, in wording that leaves no room, that «every page you care about should have a link from at least one other page on your site».

Read that again. At least one. The bar cannot go any lower. And yet, on my own site, those four pages cleared it only thanks to the homepage. Between them, not a single link.

It also says something we keep forgetting: Google can crawl your link only if it is a proper link in the page’s code (an <a> element with an href attribute). A box that reacts to a click and changes page in code is, to a crawler, a box. I have seen it in menus that «work fine», and I wrote it up in detail in why your page can be empty as far as the crawler is concerned.

The link text is not decoration

Let us leave the where and look at what we write inside the link, because that is where most of the waste happens.

The same documentation asks for link text that is «descriptive, reasonably concise, and relevant to the page that it’s on and to the page it links to». And it explicitly discourages «click here» and «read more».

Think about it practically. The reader scans the page with their eyes. The blue words are the first thing they catch. If every blue word on your page says «here», «here» and «more», you have handed them three unlabelled buttons and asked them to choose. They will not choose. They will leave.

And a warning from the opposite direction: do not overcorrect. Writing the same target keyword phrase into every link, stuffed with the terms you are chasing, is not clever. Google calls it keyword stuffing in link text and files it under its spam policies. There is a road between «here» and «cheap web design agency London», and it is wide enough for all of us.

The «related articles» box at the end of the page is useful and it should exist. But let us be honest about what it catches: it catches only the people who reached the end. The minority.

The link that catches the majority is the one sitting in the third paragraph, right where you have just used a word the reader does not own yet and would like to understand without opening a new search. That is where you keep them. Not because you trapped them, but because you answered the question in the moment it appeared.

My own rule, and I do not claim it is law: every time I write a sentence I would like to explain but explaining would derail the article, that is where a link belongs. If the page it should point to does not exist, I note it down and it becomes the next article. That way the blog grows a shape instead of just growing longer.

Because the question will come up: does all of this only apply to classic search?

No. The crawlers behind answer engines walk your site the way crawlers have always walked it, by following links. I see it in my own server logs, and I described it in which answer engine crawlers visit your site. A page nobody links to does not get luckier there, it gets unluckier, because in an answer engine there is no second page of results for someone to find you out of curiosity. There is one answer, and it either includes you or it does not.

The other thing that changes: an answer engine is not only judging a page, it is building a picture of who you are. And that picture is far easier to build when your pages point at each other with descriptive text, because you are telling it yourself that these ten pieces belong to the same thing.

How you know it is working

Let us not leave this at «it feels better». Two things are measurable, and both exist in any serious analytics tool:

Exit pages, meaning where people leave from. If a page sits permanently near the top of that list, its content is not always to blame. Very often it simply has nowhere to send anyone.

Pages per visit. It is a blunt number and it tells you nothing on its own, but its movement over time tells you something. If it does not budge at all in two months after serious linking work, something went wrong, and it is usually the link text.

And one measurement no analytics tool will give you, because it cannot know it: how many of your pages no other page links to. That one only comes from crawling your own site. Take the addresses from your sitemap (the list of all your pages), collect every link on every page, and subtract. The number will surprise you, and I do not say that to frighten you. I say it because it surprised me.

Why this is not a Sunday afternoon job

I want to be blunt here, because the opposite would be convenient for me and useless for you.

There is no version of this where you sit down one afternoon, link your eighty pieces, and the matter is closed. It is not a question of effort, it is the nature of the thing:

Every new article you publish is born an orphan. No older piece knows about it, because they were written before it existed. If you do not go back and add two or three links pointing at it, it will hang there alone forever.

Every redesign, every address change, every «let us tidy up the menu» breaks links that used to work. Quietly. Nobody notifies you, and the visitor who lands on the broken one will not tell you either, they will just leave.

And every time what you sell changes, which page should lead to which changes with it. It happened to me this week with four pages that were not speaking to each other, and I build the site myself. If it slips past me, imagine how easily it slips past you.

So the honest answer is not «do it now and do it again in three months». It is that this needs continuous checking, the same kind you run for availability or for backups. Someone has to look, every time something changes, at which page has been orphaned and which link has broken.

And so the wall becomes a corridor

Let us go back to the reader from the beginning, the one who read for seven minutes and found a footer.

They did not leave because they were bored. They left because the road ran out.

And the road is not built by bolting a «related articles» box onto the footer in a hurry. It is built one sentence at a time: every time you write something the reader will want to chase further, you hand them the link yourself, with text that says what they will find there. And then, every time something on the site changes, someone checks which of those links survived.

Because the reader who made it to the end already did the hard part. The easy part, telling them where to go next, is worth not losing.

If you would like to see how many of your pages are hanging on their own right now, that is something we check: we find the pages no link points to, the links that have broken and the places where the reader should be led towards your services. What you get is a list of what we found and what we recommend, in order of priority. Send us your site address through the form. This work is part of our SEO and GEO service.

The whole picture: this article covers one part of a larger subject. If you are after the whole of it, from beginning to end, it is in the guide to answer engines.