What Is an Lsi Anchor? A Plain-English Explanation

An LSI anchor is a link whose text uses a synonym or related phrase instead of the exact keyword. Google does not use LSI, so the term is best read as a related-phrase anchor.
An LSI anchor is a link whose clickable text uses a word or phrase related to your target keyword, such as a synonym or a close variation, instead of the exact keyword itself. The term comes from latent semantic indexing, a 1980s retrieval method Google does not use, so read it as a related-phrase anchor.
Key takeaways
- Google has said LSI keywords do not exist as a ranking concept, and its ranking-systems guide names BERT, RankBrain and neural matching instead of LSI.
- An LSI anchor is best read as a related-phrase anchor: a synonym or variation of the target phrase, such as "SEO software" linking to a page about "SEO tools".
- Ahrefs studied 384,614 pages and found only a weak link between exact-match anchors and rankings, with a Spearman correlation of 0.14.
- Google lists links with optimized anchor text in guest posts and press releases as link spam, and breaking its spam policies can cost a site its rankings.
- A natural anchor mix leans on branded, generic and related-phrase anchors, and you can check yours in a backlink tool's Anchors report.
What an LSI anchor means
An LSI anchor is a hyperlink whose visible text is a term related to the target keyword, rather than the keyword itself.
LSI stands for latent semantic indexing, a retrieval method from the late 1980s that grouped words by how often they appeared together in documents. It used a math step called singular value decomposition to find hidden links between terms, as MarTech explains.
SEO guides later borrowed the label for any related word used as anchor text, such as "link building guide" pointing to a page about "backlink strategies". That borrowing is where the confusion starts, because the label sounds like a Google system and it is not.
The idea sits inside the wider topic of anchor text, which is the clickable text of any link. An LSI anchor is one flavor of it, next to exact-match, partial-match and branded anchors.
How an LSI anchor works
An LSI anchor works the same way as any anchor: it tells readers and Google what the linked page is about, using a related phrase instead of the exact target phrase.
Google says good anchor text is descriptive, reasonably concise and relevant to both the page it sits on and the page it links to. It also says the words before and after a link matter, so the sentence around an LSI anchor adds context too, according to Google's link best practices.
The old LSI method itself worked by counting word patterns across a fixed set of documents. Google does not run that method on the web.
Its guide to ranking systems lists no LSI system among the systems it names, as of November 2026. The same guide says Google's systems look at signals across hundreds of billions of web pages, according to Google Search Central.
What Google does use instead is a set of language systems, including BERT, RankBrain and neural matching. BERT began rolling out in October 2019 for English queries and first affected about 1 in 10 searches, according to Search Engine Land in 2019.
RankBrain links words to concepts, so a page can rank even when it does not contain every exact word in a search. That is the real mechanism behind related-phrase anchors: meaning, not a secret LSI word list.
Google launched its Knowledge Graph in 2012, more than 20 years after LSI appeared. It connects entities such as people, places and things, which is a different job from counting word patterns, as MarTech explains.
Why it matters for rankings and AI search
It matters because varied, descriptive anchors help Google understand a page without looking engineered, while a wall of identical keyword anchors is a known spam signal.
Ahrefs studied the anchors pointing to 384,614 pages that ranked in the top 20 for 19,840 keywords. It found a Spearman correlation of 0.14 between the share of exact-match anchors and ranking position, which it called relatively weak, in a study published in 2019, according to Ahrefs.
Ahrefs ran a second test on 16,000 pages with random anchors and found the exact-match keyword in the words around a link had a Spearman correlation of 0.06 with rankings, which it called no notable effect.
The same study found the median share of exact-match anchors was zero at every ranking position. In plain terms, many top-ranking pages had no exact-match backlinks at all.
For AI search, the link profile still counts as a trust signal. SE Ranking analysed 129,000 domains and 216,524 pages across 20 niches in 2025 and found sites with over 32,000 referring domains were 3.5 times more likely to be cited by ChatGPT than sites with up to 200, according to SE Ranking.
Ahrefs also found in a 2026 study of 1.4 million ChatGPT prompts that ChatGPT cited 88% of the pages it pulled from ordinary web search. A clear, descriptive anchor helps a page get found in that search step, even if no study proves anchors alone win citations, according to Ahrefs.
Real examples of LSI anchors
Real LSI anchors are easy to spot: the link text is a synonym, a variation or a longer phrase built around the target topic.
- An exact-match anchor repeats the target phrase word for word, such as "best CRM software" linking to a page that targets "best CRM software".
- An LSI anchor uses a synonym, such as "customer management tools" linking to that same CRM page.
- An LSI anchor uses a longer related phrase, such as "software for tracking sales leads" linking to that same CRM page.
- A partial-match anchor contains all the words of the target phrase plus extra words, such as "affordable CRM software for small teams".
- A branded anchor uses the brand name itself, such as "Salesforce" linking to a Salesforce page.
- A generic anchor says nothing about the topic, such as "click here" linking to the CRM page.
| Anchor text | Anchor type | What it tells Google |
|---|---|---|
| best CRM software | Exact match | The page targets that exact phrase |
| customer management tools | LSI anchor | The page covers a close synonym topic |
| software for tracking sales leads | LSI anchor | The page covers a longer related phrase |
| affordable CRM software for small teams | Partial match | The page covers the phrase with extra words |
| Salesforce | Branded | The page belongs to that brand |
| click here | Generic | Nothing about the topic |
The LSI rows describe the topic without repeating the exact keyword.
How to check and measure your anchor mix
You check your anchor mix by exporting the Anchors report in a backlink tool and counting how often each anchor type appears.
- Open your backlink tool, enter your domain and open the Anchors report, which lists every anchor pointing to your site.
- Export the list and sort it by the number of referring domains, not by raw link count, so one site with 50 links does not skew the result.
- Label each anchor as exact-match, partial-match, related-phrase, branded, generic or naked URL, then work out each group as a share of the total.
- Flag any page where one keyword anchor dominates, then check its rankings and traffic in Google Search Console before you change anything.
- Repeat the check every 3 months, or after a big link campaign, so a sudden spike shows up early.
Ahrefs notes that most SEOs recommend keeping exact-match anchors near 1% to 5% of a profile. Google publishes no safe percentage for any anchor type, as of November 2026, so treat that range as a habit, not a rule, according to Ahrefs.
The same Ahrefs study of 384,614 pages found the median exact-match share was zero at every ranking position. That does not set a safe limit, but it shows top pages rarely depend on exact-match anchors.
Good vs bad LSI anchors and the risks
A good LSI anchor describes the linked page in words a reader would use, while a bad one hides the topic or repeats a sales phrase.
Good LSI anchors share three traits:
- They name the topic, such as "email marketing guide" for a page about email marketing.
- They fit the sentence, so a reader can guess the destination before clicking.
- They vary across links, so ten links to one page do not all use the same phrase.
Bad LSI anchors share the opposite traits:
- They say nothing about the page, such as "this" or "here".
- They stuff in extra keywords, such as "best cheap fast CRM software deals 2026".
- They appear in paid guest posts or press releases with the same keyword phrase on many sites.
Google's spam policies list links with optimized anchor text in articles, guest posts or press releases distributed on other sites as link spam. Sites that break the policies may rank lower or not appear in results at all, as of November 2026, according to Google Search Central.
Penguin, the system built to fight link spam, was announced in 2012 and folded into Google's core ranking systems in 2016. It is one reason heavy exact-match anchor use has carried risk for over a decade, as Google's ranking-systems guide records, and our guide to Google Penguin explains the history.
Common mistakes and myths
The most common mistake is treating LSI as a Google system and buying lists of LSI keywords to sprinkle into anchors.
- Myth: Google uses LSI to rank pages. John Mueller said in 2019: "There's no such thing as LSI keywords." Google's ranking-systems guide names BERT, RankBrain and neural matching, not LSI, as MarTech explains.
- Mistake: repeating one related phrase in every link. Ten links with the same anchor look planned, even when the phrase is a synonym.
- Mistake: chasing a fixed anchor ratio. No Google page publishes a safe percentage, so a ratio copied from a guide is a guess, not a rule.
- Mistake: paying for keyword-rich anchors in guest posts. Google lists that pattern as link spam, and the risk sits with the site that receives the links, as our guide to the Google link spam update explains.
- Myth: more related words always help. Relevance comes from the page and the words around the link, not from the number of synonyms in the anchor.
Frequently Asked Questions
Is an LSI anchor a Google ranking factor?
No. LSI is a retrieval method from the late 1980s, and Google has said it does not use LSI keywords. What helps rankings is a clear, descriptive anchor that matches the linked page, plus links from trusted pages. Treat LSI anchor as shorthand for a related-phrase anchor, not as a factor Google measures.
What is the difference between an LSI anchor and a partial-match anchor?
A partial-match anchor contains all the words of the target phrase plus extra words, such as "affordable CRM software for small teams". An LSI anchor uses a related phrase instead, such as "customer management tools". The two overlap in practice, and both are safer than repeating one exact-match phrase everywhere.
How many exact-match anchors are safe?
Google publishes no safe number or percentage, as of November 2026. Ahrefs found that the median exact-match share was zero at every ranking position across 384,614 pages. Use that as a warning, not a target: varied anchors, with branded and related-phrase links in the mix, look more natural than a fixed ratio.
Can LSI anchors help in AI search?
No study shows that related-phrase anchors alone win AI citations. What the data does show is that strong link profiles matter: SE Ranking found in 2025 that sites with over 32,000 referring domains were 3.5 times more likely to be cited by ChatGPT. Descriptive anchors support that profile by making links easy for search systems to read.
Should I ask other sites to use LSI anchors when they link to me?
No, do not script anchors for other sites. You rarely control how others link, and asking for keyword-rich anchors in guest posts is the pattern Google calls link spam. Earn links with useful pages, and let other writers describe them in their own words.
Sources
- Google Search Central - SEO Link Best Practices for Google (accessed November 2026)
- Google Search Central - Spam Policies for Google Web Search (accessed November 2026)
- Google Search Central - A Guide to Google Search Ranking Systems (accessed November 2026)
- Ahrefs - What Is Anchor Text? (2019, updated 2026)
- MarTech - Latent semantic indexing: Why marketers do not need to worry about it (accessed November 2026)
- Search Engine Land - FAQ: All about the BERT algorithm in Google search (2019)
- SE Ranking - How to Optimize for ChatGPT (2025)
- Ahrefs - Where Does ChatGPT Get Its Information? (2026)
The BacklinkPrice Team researches backlink prices, publisher sites and link building tools. We check prices and plans against each provider's own pages and date them, so you know how current the numbers are.
More from BacklinkPrice TeamRelated guides





