Search for SEO writing tips and you will soon find advice to sprinkle "LSI keywords" into your content. Tools promise to generate them, checklists demand a quota of them, and some guides claim Google uses Latent Semantic Indexing to judge whether your page is relevant.
The problem is that Google has said, plainly, that there is no such thing as LSI keywords. The underlying technology is a decades-old information retrieval method that was never designed for a web-scale search engine. Yet the advice behind the myth, use related words and cover a topic properly, is not wrong. It is just explained badly.
This article separates the myth from the reality: what LSI actually is, what Google has said, why related terms still matter, and a practical process for covering topics in a way modern search understands.
Key Takeaways
- Google's John Mueller said in 2019 that there is no such thing as LSI keywords.
- Latent Semantic Indexing is a 1980s retrieval technique, patented in 1989, built for small, static document collections.
- "LSI keywords" in SEO tools are usually just related or co-occurring terms; the label is misleading.
- Google uses far more advanced language systems, such as BERT, to understand meaning.
- Cover the topic thoroughly and naturally: answer sub-questions, use the terms experts and users use, and avoid keyword lists.
What Google Has Said
The clearest statement came from Google Search Advocate John Mueller on X (then Twitter) on 30 July 2019, reported the next day by Search Engine Roundtable.
“There's no such thing as LSI keywords -- anyone who's telling you otherwise is mistaken, sorry.”
— John Mueller, Google, as reported by Search Engine Roundtable, Google: There Is No Such Thing As LSI Keywords
Google's own documentation never mentions LSI. Instead, its SEO Starter Guide encourages you to think about the words a user might search for, noting that people who know a lot about a topic may use different words from beginners. That is a point about language and audience, not about a hidden semantic index.
Where the LSI Idea Came From
Latent Semantic Indexing (also called Latent Semantic Analysis) is a mathematical technique developed in the 1980s to find relationships between terms and documents in a collection. A patent for it was granted to Bell Communications Research in 1989. It helped early retrieval systems match documents that used different words for the same concept.
The late Bill Slawski, who studied search patents for years, was a persistent critic of the SEO use of the term. Ahrefs quotes him explaining that LSI technology was not created for anything the size of the web and that the label confuses people.
“Calling it LSI is misleading people.”
— Bill Slawski, quoted by Ahrefs, LSI Keywords: What are They and Do They Matter?
| Claim | Reality |
|---|---|
| Google uses LSI to rank pages | Google has said there is no such thing as LSI keywords, and its documentation never refers to LSI |
| You need a set number of LSI keywords | No quota exists; Google has no preferred keyword counts or word counts |
| LSI tools reveal Google's semantic model | Most tools list related or co-occurring terms from search results or autocomplete |
| Synonyms are LSI keywords | Synonyms are just synonyms; Google handles them with its own language systems |
| Related terms are useless | Covering related sub-topics naturally makes content more complete and useful |
How Google Actually Understands Language
Google has invested heavily in language understanding. When it introduced BERT in Search in 2019, Google said it would help Search better understand one in 10 searches in the U.S. in English, especially longer, conversational queries where small words like "for" and "to" change the meaning. Since then, Google has layered on more advanced AI systems that power features like AI Overviews and AI Mode.
The practical lesson: Google does not need you to feed it a list of related words. It needs a page that genuinely covers the subject, uses clear language and answers what searchers want. Our guides to semantic SEO and entity SEO explain how this works in more depth, and RankBrain covers an earlier machine-learning system.
What to Do Instead of "LSI Keywords"
The useful part of the LSI advice is "cover the topic, not just the phrase". Here is a better way to do it.
- Start with search intent. Work out what the searcher wants: a definition, a comparison, a how-to or a product. Our search intent guide explains the types.
- List the sub-questions. Use People Also Ask, autocomplete, related searches and your own customer questions.
- Study what top results cover. Note sub-topics that most good pages include, then look for what they all miss.
- Gather the vocabulary of the field. Talk to experts, read product documentation and listen to how customers describe problems.
- Outline the page around those topics. Use descriptive headings that map to sub-questions. A content brief helps writers stay on track.
- Write naturally. Related terms will appear on their own when you explain a topic properly.
- Review for gaps, not counts. Ask whether a reader would leave with unanswered questions.
Related terms: helpful vs unhelpful use
| Topic: "how to descale a kettle" | Helpful | Unhelpful |
|---|---|---|
| Related terms | Limescale, white vinegar, citric acid, hard water, rinse | "kettle descaling, descale kettle, descaling kettle tips" repeated |
| How they appear | Inside clear steps and explanations | In a list at the bottom or forced into sentences |
| Sub-questions covered | How often to descale, safety, removing the vinegar taste | None; the same point restated |
Are "LSI Keyword Tools" Useless?
Not entirely. Many tools that use the LSI label simply surface related searches, co-occurring terms in top results or autocomplete suggestions. Those can be useful prompts for sub-topics you might have missed. Just ignore the label, ignore any quota, and never paste the list into your content. Our guides to keyword research and types of keywords cover better ways to find related terms.
Common Mistakes
- Treating LSI as a ranking factor. Google has said LSI keywords do not exist.
- Keyword lists in content. Dumping related terms into a paragraph reads badly and can resemble keyword stuffing.
- Chasing tool scores. Content scores based on term counts are proxies, not Google's judgement.
- Ignoring expertise. Real expertise naturally produces the right vocabulary; templates do not.
- Thin pages "optimised" with terms. Adding words to a page that adds no value does not fix it. See thin content.
Related Guides
- Negative SEO: How to Spot and Protect Against It
- The Google Sandbox: Myth or Reality?
- How to Write an SEO Proposal (With Template)
Frequently Asked Questions
Does Google use LSI keywords?
No. Google's John Mueller said in 2019 that there is no such thing as LSI keywords, and Google's documentation does not mention LSI.
What do SEO tools mean by LSI keywords?
Usually related or co-occurring terms taken from search results, autocomplete or related searches. They can suggest sub-topics, but the LSI label is inaccurate.
Should I still use related keywords?
Yes, naturally. Covering a topic thoroughly means using the vocabulary and answering the sub-questions that real users and experts use.
What is latent semantic indexing?
A mathematical information retrieval technique from the 1980s, patented in 1989, that finds relationships between terms and documents in a fixed collection.
What replaced LSI in Google?
Google never used LSI as described by SEO tools. It relies on its own language-understanding systems, including BERT, introduced in Search in 2019.
Conclusion
LSI keywords are a myth, but the instinct behind them, cover topics properly in natural language, is sound. Focus on intent, sub-questions and expert vocabulary rather than term lists and quotas. If you want content that covers topics in depth and ranks on its merits, our content writing and SEO teams can help. Get a free quote.
References
- Search Engine Roundtable: Google: There Is No Such Thing As LSI Keywords
- Ahrefs: LSI Keywords: What are They and Do They Matter?
- Google Blog: Understanding searches better than ever before (BERT)
- Google Search Central: SEO Starter Guide
- Google Search Central: Creating helpful, reliable, people-first content
- Semrush: What Are LSI Keywords & Why They Don't Matter
