Creative agency glossarySEO and Analytics

Keyword Stemming

Keyword stemming is the way search engines reduce words to their root so that a search for one form of a word also matches its plurals, tenses and variations.

What is keyword stemming?

Keyword stemming is a search engine recognizing that film, films, filmed and filming are the same idea. The engine strips a word back to its stem and matches on that, so a page that uses one form can rank for a search that uses another. Google has done this since the early 2000s, and its language systems have since gone well beyond stems to synonyms, misspellings and meaning, but stemming is the foundation: the point at which a search stopped being an exact string match.

In SEO conversation the term is used two ways. Strictly, it is what the engine does. Loosely, it means writing content that uses the natural variations of a keyword rather than repeating one form, in order to match the range of ways people search.

Why it matters

Why keyword stemming matters

Because it removes a whole category of pointless work. Before stemming, SEOs wrote separate pages or stuffed separate variants for singular, plural and verb forms; now one well-written page covers them all, and trying to game each variant separately produces exactly the kind of repetitive copy that reads badly and ranks worse. Understanding stemming is permission to write normally.

It also matters for research. Keyword tools report volumes per exact phrase, and a plan that treats brand film and brand films as separate targets double counts the demand and splits the effort. Grouped, they are one cluster with one page.

In practice

How keyword stemming works in practice

  1. 01

    The engine side is invisible. A query comes in, each word is reduced to its root and expanded to its known variants and synonyms, and the index is searched for all of them, with closer matches usually scoring higher. Google highlights the variants in bold in results, which is the easiest way to see stemming at work: search for one form and watch the engine bold another.

  2. 02

    The writer side is simply good prose. Use the word the way a person would in each sentence, which naturally produces the stems and variants; cover the related phrasings that mean the same thing; and stop counting occurrences. Where a brand name or a technical term has a specific form that must be matched exactly, use that form in the title and H1 and let the body vary. Keyword tools that cluster by stem or by results overlap do the grouping for research.

Watch out

What people get wrong about keyword stemming

Assuming it is complete. Stemming handles regular variations well, but irregular forms, compound words and terms with different meanings in different forms can slip through, so the primary keyword still belongs, in its exact form, in the places that matter most. The other mistake is the reverse: using stemming as a reason to repeat every variation in a paragraph, which is keyword stuffing with a different name. The third is confusing stemming with synonyms; the engine handles both, but a page about film production is not automatically a page about video production, and the content has to say so if it is.

Questions

People also ask

Does Google use keyword stemming?

Yes, and has for about two decades. A search for one form of a word returns pages that use its plurals, tenses and related forms, and Google bolds those variants in the results. Its current language models go well beyond stems to synonyms and meaning.

Should I include every variation of a keyword in my content?

No. Write naturally and the variations will appear where they belong; deliberately listing every form reads as stuffing. Put the exact primary phrase in the title and main heading, and let the body use the language a person would.

What is the difference between stemming and LSI keywords?

Stemming groups forms of the same word, such as design, designs and designed. LSI keywords, as SEOs use the term, are different words that tend to appear in the same context, such as designer, typography and layout. The engine uses both to understand what a page is about.

Next to it in the glossary

  1. Keyword MappingKeyword mapping is the process of assigning each target keyword or cluster of keywords to one specific page on a site, so every search has a home and no two pages compete for it.
  2. Keyword RankingKeyword ranking is the position a page holds in the search results for a particular term, and tracking it over time is how SEO progress is measured.
  3. Keyword ResearchKeyword research is the process of finding out what people type into search engines about a subject, how often, and with what intent, so that content and campaigns can be built around real demand.
  4. Keyword StuffingKeyword stuffing is the practice of loading a page with a search term far more often than natural writing would, in an attempt to rank for it, and search engines have penalized it for years.
  5. KeywordsKeywords are the words and phrases people type into a search engine, and the terms a page is written to be found for.
  6. KPI DevelopmentKPI development is the process of choosing, defining and agreeing the small set of measures that will tell a team whether its work is achieving the business goal.

Every SEO and Analytics term →

In the work

Where it shows up

All our creations →

From the journal

Read around it

The journal →

Talking the talk is one thing

We make the work.

RIOT is a creative agency and production studio in New York. Strategy, brand, film, photography, design and the campaigns that carry them, made by one team. A few lines is enough to start: we read every brief the same day.

    Footage