Internal linking: who refers to whom and who is left without links
Internal linking of the site is work with how your pages transfer weight to each other: how many links lead to each address, from which pages they come and with what text they are signed. It does not start with an arrangement of links, but with two lists. The first is all your addresses. The second is addresses to which at least one link leads. The difference between them is the orphans, which the bot reaches last or does not reach at all.
We build the link graph for free and without access. Next - 10-20 working days for the scheme and implementation, and we name the amount after we see the numbers. Immediately the limit: the weight is redistributed between your same pages, its sum does not grow, so this work does not promise an increase in items.
graph of internal links, search for orphans, measurement of click depth, breeding anchors, link blocks in the template
after a free audit
Price
30
Guarantee
days after project sign-off
graph of internal links, search for orphans, measurement of click depth, breeding anchors, link blocks in the template
What are we doing?
10-20
Term of works
working days - the deadline is determined by the number of orphans and the place where the links live, not the size of the directory
free analysis of the graph, 2-4 working days, you need the website address and a list of your addresses
Before the estimate
40
Measured depth
categories on four levels — 5 root, 16, 10 and 9; nine sections for four transitions from the main one
the internal link leads to the address that gives the transition or is a copy of the product
The most frequent find
increase in items - links within the site transfer weight, not create it
What we do not promise
Process steps
Transparent stages with approval at every step
Total duration:10–20 days
1
Traversal and graph construction
2-4 working days
We go through the site and for each address we record how many internal links lead to it, from which pages and with what text. A separate list is source links, including those hidden by styles. The output is a table, which can be further argued objectively.
2
Comparison with a complete list of addresses
1-2 working days
We take your list of all addresses - a site map or download from the database - and subtract the crawl result from it. What remains are orphans. Here we count the depth of the click: how many transitions from the main page to each type of page.
3
Scheme: who transfers the weight to whom
2-3 working days
We divide the site into levels and decide which pages collect weight and which give away. You name the priority sections, we show how many links each one has now and how many it takes to stop being equal to the delivery page.
4
Implementation in a template
3-8 working days
We put breadcrumbs with markings, blocks of adjacent items in the card and category, links of articles with the catalog. At the same time, we clean the links leading through the transition or to a copy of the address - otherwise, the new blocks will simply add jumps to the already existing ones.
5
Anchors and rules for content
1-2 working days
We divide the texts of links to priority pages and write down the rules for those who will fill the site further: where to link to a new page, what anchors to take, how many to put in the text. The rules are short, otherwise they are not used.
6
Control bypass and transfer
1 working day
We repeat the first step with the same tool and compare two numbers: orphans and click depth before and after. We set an event for transitions to new blocks, so that in a month we can see whether they are being used.
Technologies & integrations
What we build on and what it connects to
Stack
crawling the site with the construction of a link graph - counts incoming links to each address and saves the source page together with the anchor text
comparison of the bypass with a complete list of addresses — site map or download from the database; it is impossible to find an orphan without a second list
Screaming Frog — inbound and outbound links and click depth reports; the free version is limited to five hundred addresses
Search Console, internal links report - shows which pages Google itself considers the most linked; data comes with a delay
CMS templates — blocks of adjacent items and breadcrumbs live in the template, not in the text of the page, so they are edited once for the entire site
BreadcrumbList markup is a path to the page that the search engine reads in a structured way, rather than guessing from the layout
GA4 — transitions to new blocks of links; the counter doesn't improve anything, it just lets you see if the blocks are being used
Integrations
Google Search Console
GA4
Google Tag Manager
Bing Webmaster Tools
What's included
Complete list of work and what you get as a result
Graph of internal links: for each address - how many links lead to it and from which pages
List of orphan pages: the difference between a complete list of your addresses and the result of a crawl - with one crawl, such pages are not found in principle
Click depth measurement by page type with a list of what lies deeper than the third transition
Analyzing anchors: how many different texts lead to each target page, where there are massive identical ones and where there are meaningless "here" and "read"
A list of internal links that lead to a transition, to an error, or to a copy of the address of the same product
Audit of outgoing links: where the site itself leads, including what the author has sewn into the theme and what is not visible on the screen
Hub scheme: which pages collect weight and to whom they distribute it
Implementation of blocks in the template: breadcrumbs with BreadcrumbList marking, selections of adjacent items in the card, connection of blog articles with catalog categories
The rules of anchors are in writing - from where to link to a new page and with what texts
An event in analytics for transitions to new blocks — otherwise it is not known whether they are used
Post-implementation audit trail with two numbers: how many orphans are left and how the click depth has changed
When this service isn't right
What's not included — so there are no surprises at delivery
Buying links from other people's sites is a different job and a different budget
Writing texts for pages to which new links lead
Reorganization of the directory structure and renaming of categories
Automatic distribution of links without rules: a module that links anything to anything, we do not put
Working with links that lead to your site from other people's resources
Who it's for
Situations where this service delivers results
Scenario 1 of 5
The catalog is growing, and new positions do not appear in the search for a long time
The product is installed, it opens at the direct address, but only the one who knows the way gets there. The most common reason is simple: there is one link to the card from the category page, and that category itself is on the third level. The bot reaches her last. We count how many such pages you have and how many of them are without incoming links.
We'll review your situation in a free audit
Scenario 2 of 5
There are pages that you consider to be the main ones, but the site does not show this
A selection for the season, landing for a popular combination of filters - for a search engine, they are equivalent to a page about the return conditions, if there are equally few links to them. Site priorities are read from the link graph, not from your intentions.
Scenario 3 of 5
Navigation is maintained on several parallel layers
Categories, brands, tags, selections — each layer leads to the same products in its own way. In the measured underwear store, there are two such layers: 129 categories and 334 brands with 115,456 product items. The question is not whether both are needed, but whether they lead to the same product address.
Scenario 4 of 5
The catalog is expanded in depth through a small nomenclature
When couplings and tees differ in diameter and thread, without four levels of nesting, the buyer will not find the desired part. The price of such a structure is pages for four transitions from the main page. In the store of irrigation systems, we measured exactly this: 5 root categories, 16 of the second level, 10 of the third and 9 of the fourth.
Scenario 5 of 5
The site is built on a purchased theme, and no one has checked what it does with links
In the live store, we found four hidden third-party domain links with display:none style in the template. The owner did not know about them: these are the source links of his site, and they were placed by the author of the topic.
Scheme for your priorities against the "similar goods" box module
How this option differs from the alternative
Auto-selection module in CMSOur approach
What determines the choice of linksmatch by category, price, or a random selection from the same sectionrules for sections that you have named priority
What address does the link lead to?to the address at which the goods are currently opened - hence the copies under different pathsto the product's canonical address, one for all paths
Link textthe name of the product is the same in all places of the sitevaries: full name, abbreviated, name with description
Pages without incoming linksthe module does not know that such a page exists: it works inside a categoryare counted separately and closed by address
What's next?works by itself and for free, until the template was changedthe rules should be reviewed after each major structural change
Conclusiona normal option for an equal directory without expressed prioritiesworth where there are priority partitions and orphans
Free analysis of the graph of internal links
The graph is built by traversal, and the site address is enough for it. The output shows how many internal links lead to each address, from where exactly, and which addresses do not have any. To count orphans, you will additionally need your list of addresses: site map or download from the database.
What we measure
Pages without any incoming linkThey are calculated by comparing two lists: a complete list of your addresses and a list of where at least one link leads. It is impossible to see the orphan by the detour itself - the detour follows the links, but there are none for it.
Click depthHow many transitions from the main page to each page. Separately collect everything that lies deeper than the third transition.
Distribution of anchorsWhat texts are in links to the same page. We show mass identical and meaningless ones like "here" or "in more detail".
Links leading past the jumpInternal links to addresses that return a transition or error. The most frequent finding of this analysis.
The target of the link is a copy or an originalDo links to one product lead to the same address, or to multiple copies of it under different category paths.
Outbound template linksWhere the site itself leads, including what the theme author sewed into the markup. We found hidden links in the live store.
What you get
A table of addresses with the number of incoming internal links and a list of source pages.
A separate list of pages without any incoming link.
Distribution of click depth by page types.
A list of links leading to a transition, an error, or a copy of an address.
A conversation about findings, where conclusions can be disputed.
Timeline: 2-4 working days
Why is it free
The search is automated, and the result immediately shows whether there is work for a few days or a few weeks. Negotiating an amount up to this figure is guesswork.
What's next
Next, there is a scheme: which pages collect weight, which ones are distributed, where the links to priority sections come from. The implementation is divided into blocks, and you can opt out of any of them.
Short form: your contact and site URL
Did not find your case?
Describe how it works on your side — we will tell you whether “Internal relinking” fits and what it means in your situation. No brief and no call: one question, one answer.
Contract, act and 30-day warranty
Every project gets a written contract: scope, deadlines, amount, acceptance procedure. After delivery — act and invoice, then 30 calendar days of warranty.
Sole proprietor & bank transfer
The contractor is a registered sole proprietor. Payment by invoice with closing documents.
Rights & access — yours
Code, design and materials transfer to you after full payment. Domain, hosting, repository and analytics are registered to you.
Client portal instead of email chains
During the project you get access to a portal: contracts, invoices, acts and project status in one place.
European clients
Among our work — projects for Norway, Bulgaria, Moldova and Spain.
Verifiable numbers
Every case in the portfolio comes with a link to a live site and a technical measurement.
Audit first, then pricing
There is no price list on the site intentionally: the scope of the same work differs multiples between clients.
We say "no" when unsure
If the task isn't ours or the deadline is unrealistic — we tell you upfront.
What affects the price
Why two seemingly identical tasks are priced differently
How many addresses are there on the siteThe graph is built on all addresses, so the traversal cost grows with the directory. But it grows slowly: the difference between a thousand addresses and ten thousand is hours of machine time, not weeks.
Why orphans became orphansHalf a thousand pages dropped due to one rule in the template are treated by editing this rule in a day. Half a thousand, which fell out piecemeal for years, are treated piecemeal. The number in the report is the same, the amount of work is different - and it is impossible to name it before analysis.
Where the links liveMenu, footer, card template are a few places in the code. Links inside product descriptions and blog articles — as many edits as there are records in the database. The second is sometimes cheaper not to touch at all, but to compensate with blocks in the template.
How many layers of navigationCategories, brands, tags, seasonal selections — each layer adds its own set of rules and its own risk that two layers will lead to different addresses for the same product. One layer is a simple scheme, four is a separate stage for reconciliation.
Do you need new pages for hubs?Sometimes there is no one to distribute the weight to: there is simply no intermediate layer of partitions. Then the work is divided into two - linking and creating pages, and we call the second one separately, because it is no longer an arrangement of links.
The pace of changes on the siteThe catalog, in which chapters appear and disappear every week, spawns new orphans continuously. The one-time scheme quickly becomes outdated, and it is more honest to count the round once a quarter.
Cases
Tasks and results in numbers — all metrics measured by us
store of automatic and drip irrigation systems, catalog of 300 items
Task
See how the catalog navigation is arranged from the small nomenclature.
Solution
Traversal with construction of a graph, calculation of levels of nesting, checking of source links of a topic.
Result
300 products are divided into 40 categories in four levels: 5 root, 16 second, 10 third and 9 fourth — nine sections lie four transitions from the main one. The only working address list is an HTML map of 89 links with approximately 340 addresses including categories. Two of the five root categories are empty — the links lead to a section with no products. Four hidden third-party domain links with display:none style were found in the template. Measurement 01.08.2026.
underwear retail store, catalog of over 110,000 items
Task
Check whether the weight reaches the product pages, where size and color are separate items.
Solution
Link graph on two layers of navigation and selective verification of response codes of target addresses.
Result
Navigation is based on 129 categories and 334 brands with 115,456 product items. In a sample of 150 random product sitemap addresses, 50 gave a transition, and all 50 led to a different position - a different size or color. That is, a third of the verified links do not lead to where the anchor promises. The site map does not cover about 2.8 thousand products that the store shows in its own search. Measurement 01.08.2026.
online store of professional tools, catalog of over 6,000 items
Task
Understand why internal links do not gather at the same product address.
Solution
Analysis of the ways in which the product is available from different branches of the catalog.
Result
6,486 product pages and 361 categories, and the site map has 25,537 records for 6,857 unique addresses — each product is present on average 3.9 times under different category paths. Each path is a separate address with its own link, so the weight of the product is divided into four copies. Measured on 07/31/2026.
online store of the official seller of garden tools
Task
Compare how the navigation in the directory is laid out for one and a half thousand addresses.
Solution
Analysis of site map addresses by nesting levels and verification of path marking.
Result
1,300 unique addresses: 1,231 on the first level, 64 subcategories, 5 service paths. Almost the entire catalog of 1,213 cards lies in one layer with 70 categories - the weight is distributed evenly, no group is singled out. There are 4,236 entries in the map for the same 1,300 addresses and no marking of breadcrumbs. Measured on 07/31/2026.
What we need from you
We can't start without this — best to prepare in advance
1The site address - the link graph is built without accesses, and the free part of the work requires exactly that.
2Complete list of addresses: site map, download from CMS or database. Without it, orphans do not count, because there is nothing to compare them with.
3Access templates to add link blocks to your menu, product card, and category page.
4Your list of priority sections. Weight is distributed according to your priorities: we know how to transfer it, but we do not know which product is more profitable for you to sell.
5A person who will tell you which selections and categories you plan to close in the near future — so that you don't link to them a week before closing.
If something is missing — let us know, we'll help you gather it or do it as a separate task.
FAQ
Most frequently asked questions — with concrete answers
What exactly does relinking provide if it does not add new pages?
It changes how weight is distributed between the pages you already have. At the same time, the amount does not grow: links within the site transfer it from one place to another, rather than creating it. Therefore, it would be untrue to promise an increase in items. What really changes is that a page that wasn't previously linked to becomes reachable by the bot, and the section you consider the most important no longer equals the importance of the return policy page.
How do you find pages that no one is linking to?
Two lists. The first is a complete list of your addresses: site map or download from the database. The second is the result of traversal, i.e. addresses to which at least one link leads. The difference between the lists is the orphans. By definition, they cannot be found in one round: the round goes through the links, but there are no orphans. That is why we ask for the download, and not use the site address.
How many internal links should there be on a page?
There is no exact number, and whoever names it is making it up. The benchmark is different: the links should lead to where a person might actually want to go from this page. A hundred links in the footer on every page of the site distributes weight in such a thin layer that it means almost nothing. Twenty meaningful ones from the product card — to adjacent items and to the parent category — work better than a hundred random ones.
Should the anchor be an exact key or not?
Different. If two hundred links with identical text lead to one page, it is visible from any side and reads like mechanical work. We take the name of the landing page as a basis and vary it: full name, abbreviated name, name with clarification of characteristics. We separately remove the anchors "here", "more details", "read" - they do not tell either a person or a bot what is under the link.
Do I need a related products block if I have a menu and filters?
Menus and filters distribute weight from top to bottom: the main one gives to categories, the category gives to products. At the same time, there are no horizontal connections between the goods themselves, and it is precisely in them that the bulk of the catalog lies. In the professional tool shop measured, the 6,486 product pages are in 361 categories—an average of eighteen items per section, and no links between them other than a shared category.
We have a directory spread over several levels — is that bad?
Not bad, but has a price. In the irrigation system store, we measured 40 categories at four levels: 5 root, 16 second level, 10 third and 9 fourth. The last nine are four transitions from the main one — both humans and bots are the last to reach them. The depth is justified when the nomenclature is small and without it you cannot find the necessary part. It is compensated by links bypassing the hierarchy: selections, connections between neighboring sections, bread crumbs.
Is it possible to distribute the weight through the footer - put a link to everything important there?
You can, and that's why the footer is usually overloaded. The mechanism works weakly: the link, which is on all pages of the site, differs little from the template element for the search engine. A link from a content page, where it is relevant to the topic, works stronger. We leave the footer for sections that can be accessed from anywhere: delivery, payment, contacts.
How not to break the scheme in six months, when the content manager writes new articles?
Therefore, written rules remain at the exit: from where the link to the new page should lead, which anchors to take, how many links to put in the text. The second part is the blocks in the template that work by themselves: breadcrumbs, adjacent items, the connection of the article with the category. The third is a repeat round once every six months, which counts the orphans anew. One big change in directory structure can spawn a new batch of unlinked pages in a day.
Send the site address and a list of your addresses - we will count how many pages are left without any link.
In response, a table with the number of incoming links to each address, a separate list of orphans, click depth by type of pages and a list of links leading through the jump. If the orphans don't turn up, that's what you'll hear.
From measured cases300 products are divided into 40 categories in four levels: 5 root, 16 second, 10 third and 9 fourth
There is no price list on the site on purpose: the same work differs several times over between two clients, and a “from” figure explains nothing in that case. First a free audit — we count your pages, duplicates and speed — then we name the sum and the deadline and fix both in the contract.