Off-page SEO is everything that moves a site's rankings without being on the site: links from other domains, mentions of your name without a link, listings and reviews on third-party platforms, content published elsewhere that talks about you.
In practice almost everyone reduces it to link building. Below you will find a test on a real keyword cluster that shows why that shortcut costs money, and how to measure the part it leaves out.
The four things off-page SEO covers
Keeping them separate matters, because they are bought, built and measured in different ways. Anyone who files everything under "links" ends up paying for the only one of the four that can be bought with a bank transfer.
| Off-page asset | What it is | Who decides | Where to read the number |
|---|---|---|---|
| Links from other domains | A hyperlink pointing to a page on your site | The publisher who puts it live | Unique referring domains (Ahrefs, Semrush, DataForSEO) |
| Unlinked mentions | The name of the company or product quoted without a link | Whoever writes the text | Web mentions, brand alerts |
| Presence on third-party platforms | Google Business Profile listing, marketplaces, industry directories, YouTube channel, social profiles | You, within the platform's rules | Reviews, consistency of business details, views |
| Brand demand | People searching for your name directly on Google | Nobody directly: it is the effect of the other three | Branded queries in the Search Console report |
The most neglected row is the second. A mention without a link passes no anchor text and does not show up in any backlink report, so for most SEO quotes it simply does not exist. We will see shortly why it is the asset gaining weight fastest.
The test: on the LED wall cluster, the site with fewer referring domains wins
Macropix is a Milan-based manufacturer of custom LED screens and has been one of our public projects for years: the full story, with positions before and after, is told in the guide to SEO for manufacturing. In 2020 the site did not appear for any category keyword. In July 2026 it holds a 25.55% share of voice on the Italian "ledwall" cluster, ahead of Amazon.
The interesting question is a different one: how many referring domains do the sites in that ranking have? We queried the DataForSEO backlink index on 18 September 2026 and put the two columns side by side.
| Domain | Referring domains | Share of voice, LED wall cluster |
|---|---|---|
| macropix.it | 192 | 25.55% |
| amazon.it | 99,574 | 18.08% |
| hddsvision.it | 253 | 10.17% |
| dominodisplay.com | 207 | 4.33% |
| shop.agmultivision.it | 13 | 2.28% |
| tsaled.com | 46 | 1.74% |
Amazon.it has 518 times the referring domains of macropix.it and sits behind it. HDDS Vision has 32% more and a share of voice two and a half times smaller. Domino Display has 207 and is in fourth place. If the number of domains linking to you were the deciding factor, this table would be in reverse order.
Two honest caveats, so the data does not say more than it says. The two columns were collected two months apart and from two different providers. And referring domains are a site-wide total, while share of voice is calculated on a cluster: Amazon earned its hundred thousand referring domains with the average shopping basket, not with its LED wall pages. Which is exactly the point. Overall domain authority, whatever metric you use to estimate it, is not the signal that orders a vertical SERP. We have written a separate guide on what domain authority really measures.
How much links weigh, according to people who work on the algorithm
Gary Illyes, of Google's Search team, said it on stage at PubCon on 21 September 2023: "I think they are important, but I think people overestimate the importance of links. I don't agree it's in the top three." The quote is reported by Search Engine Roundtable, which on the same occasion credits him with the example of a page ranking first for "Porsche cars" with no internal or external links.
It does not mean backlinks are useless. It means that the marginal return of the hundredth link on a site that already has ninety is much lower than link building price lists suggest.
In generated answers, mentions weigh more than links
Ahrefs compared 75,000 brands with their presence in Google's AI Overviews, calculating the Spearman correlation between citation frequency and eleven off-page metrics. The study is by Louise Linehan with data scientist Xibeijia Guan and was published on 26 May 2025.
| Factor | Correlation with AI Overview mentions |
|---|---|
| Branded web mentions | 0.664 |
| Branded anchor text | 0.527 |
| Branded search volume | 0.392 |
| Domain Rating | 0.326 |
| Referring domains | 0.295 |
| Branded traffic | 0.274 |
| Total backlinks | 0.218 |
| Ad traffic | 0.216 |
| Ad spend | 0.215 |
| URL Rating | 0.180 |
| Number of pages on the site | 0.170 |
Web mentions reach 0.664, total backlinks 0.218. Between brands in the top quartile for mentions and those in the quartile below, the jump is about tenfold: a median of 169 AI Overview citations against 14. 26% of the brands analysed did not have a single mention.
Correlation is not causation, and Ahrefs is the first to say so: a much-cited brand is also a well-known brand, and the second fact explains a large part of the first. But the order of the rows has a practical consequence that holds anyway. If you pay a trade publication for an article and it names you five times without ever linking to you, by the old yardstick you have wasted your money; by this table, you have not. We covered how the other generative engines work in the guide to generative engine optimisation.
The four levers, in order of value for money
Context matters: in most sectors your competitors have a website and a social media profile, and no structured off-page strategy at all. The gap can be built with little.
1. Listings on platforms where people already look for you
A complete Google Business Profile, identical business details everywhere, reviews requested systematically after every delivery. It is the only off-page lever you control almost entirely and the only one that produces results in weeks rather than quarters. For businesses that sell by area the return is the highest of all: start from how to optimise the listing and how local rankings work.
2. Mentions earned with data only you have
A manufacturer publishing its average lead times per process, an agency publishing the real costs of a channel, an e-commerce store publishing returns by category: these are numbers trade publications quote because they cannot find them anywhere else. They cost a day of analysis instead of a monthly budget, and they produce mentions, unprompted links and material that shows experience and expertise, all at once.
3. Content on third-party platforms
An assembly or testing video on YouTube can be indexed and cited, and it lives on a domain you do not have to defend. In manufacturing it is often the only original content the company already owns, shot for internal use and never published. We have collected the rules of the channel in the guide to YouTube SEO.
4. Links, last and on a few pages
Not because they do not count, but because they are the slowest and most expensive of the four levers, and on a site with unresolved technical issues they make no visible difference at all. The sequence we recommend is in the article on in what order to do on-page and off-page SEO.
The regional rule almost nobody mentions
In Google Search's spam policies there is an entry called site reputation abuse: hosting third-party content on a strong domain with the main aim of exploiting its ranking signals. Think of the paid "guides" or "deals" section on a national newspaper or a university portal.
The part that matters is the geographical distinction. When those pages appear in results shown to users outside the European Economic Area, which includes the UK, Google may apply a manual action to them. When they appear to users inside the EEA, the pages may be treated as separate from the main domain and rank on their own merits, without a manual action. For a UK business the conclusion is the same either way: the package bought on a national newspaper does not transfer the newspaper's strength to you. In the EEA it simply does not work; for UK searchers it can also end in a penalty. Anyone selling you that package on the strength of the host domain's authority has either not read the documentation or is counting on you not having read it.
On paid links the documentation is just as clear. Google asks for advertising links to be marked with rel="sponsored", and lists rel="nofollow" as an accepted, though not preferred, alternative, as explained on the page on how to qualify outbound links. Buying or selling links for ranking purposes without those attributes counts as a link scheme. We have put the verbatim quotes and real market prices in the article on link building; here it is enough to know that the attribute is not a formality and that nofollow does not block indexing.
How to measure off-page SEO
This is the part missing from almost every article on the subject, including those that sit firmly on page one. Without these four rows in a spreadsheet, an off-page campaign lasts as long as the budget does, and then nobody can say whether it worked.
| Metric | Where to read it | How often | What it means if it rises |
|---|---|---|---|
| New and lost referring domains | Ahrefs, Semrush, Search Console | Monthly | The net balance matters more than the total: many profiles grow and empty out at the same time |
| Web mentions of the name | Brand alerts, Brand Radar, manual search in quotation marks | Monthly | The surface a generative engine can pick you up from is growing |
| Branded queries and their clicks | Search Console query report, filtered on the company name | Quarterly | It is the most honest sign that off-page is working: people search for you by name |
| Share of voice on the cluster | Semrush, Keyword.com or an equivalent tool | Quarterly | You are gaining share over direct competitors, not just traffic in general |
The third row is the one that would settle most budget discussions. Branded queries cannot be bought or faked: if after six months of external activity nobody new is searching for the company name, the external activity has left no trace in anyone's head. We described how these readings fit into a wider plan in the article on SEO strategy, and the initial check of the link profile is part of the SEO audit.
Three mistakes we see almost every time
- Buying links to the homepage. The homepage is rarely the page that needs to rank for a category keyword. In the Macropix case, the pages holding the top three positions are product and application pages, not the homepage.
- Counting backlinks instead of referring domains. Two hundred links from a single site are worth one domain. It is the difference between the two columns every tool reports and almost no quote distinguishes.
- Starting with off-page while on-page is broken. If category pages do not exist or cannibalise each other, links arrive on a site with nowhere to put them. Start with the part that depends only on you; the shortcuts have a precise name: black hat SEO.
On-page SEO covers everything on the site that you control directly: copy, titles, URL structure, internal links, speed. Off-page SEO covers the signals that come from outside: links from other domains, mentions of your name, listings and reviews on third-party platforms, direct searches for the brand. The first can be done on your own, the second depends on how other people behave.
There is no threshold. On the Italian LED wall cluster, in September 2026, the domain with the highest share of visibility has 192 referring domains and sits ahead of amazon.it, which has 99,574. What counts is how relevant the source is to the topic and which page it links to, not the total. Gary Illyes of Google said at PubCon in September 2023 that links are not among the top three ranking factors.
Yes, and increasingly so. In the Ahrefs study of May 2025 covering 75,000 brands, branded web mentions show the highest correlation with citations in Google AI Overviews (0.664), against 0.218 for total backlinks. It is an international sample and a correlation, not proof of causation, but it suggests that an editorial mention without a link is not wasted money.
Buying or selling links for ranking purposes counts as a link scheme under Google spam policies. Advertising links are allowed if marked with rel="sponsored", or with rel="nofollow", which Google lists as an accepted but not preferred alternative. Without those attributes the practice breaks the guidelines.
It depends on the lever. Local listings and reviews have an effect within a few weeks. Links and mentions work over quarters: in the Macropix project, which started in 2020, category rankings consolidated over years, not months. The first metric to move is the number of referring domains; the last to move, and the most reliable, is the volume of searches for the company name.
Something the industry rarely says: off-page SEO is the only part of the work you cannot deliver. You can rewrite a title in the afternoon and fix a redirect before dinner, but you cannot decide that a publication will cite you on Friday. That is why off-page proposals fill up with sponsored articles: they are the only item that can be put in a calendar. Look at the calendar of a project that worked, though, and what you see is not purchases but your own publications that someone else chose to cite. The real work happens upstream, and it looks much more like choosing which data to make public than buying space.
If you want to work out which of the four levers to start from in your sector, it is the first step of our SEO projects: you can see how we set them up in the case studies we have published.
