Keyword stemming
Keyword stemming is when a search engine treats different forms of a word, like roast, roasted and roasting, as the same word.
Keyword stemming is when a search engine treats different forms of a word as one. "Roast," "roasted," "roaster" and "roasting" share a stem, so a search for "roasting coffee at home" can match a page about "how to roast coffee." The name comes from cutting a word back to its stem, its root form.
For marketers, it means you don't need to cram every form of a word into a page. Write naturally and the search engine connects the forms.
What it changed
Without stemming, a page has to contain the exact word someone typed, so writers end up squeezing in every version of it. Once search engines handled stemming, that stopped being necessary, and pages could be written for people.
Stemming was an early step. Search engines went on to match synonyms, then the meaning of a whole question. A search for "cheap espresso machine" today can find a page about "affordable espresso makers," which shares almost none of the same stems.
Stemming and AI search
AI models go much further than stemming. They read text as meaning, so word forms and synonyms mostly take care of themselves. For example, a page by Harrow & Pine that explains "how we roast our beans in small batches" can help answer "Which UK roasters do small-batch roasting?" with no need for the page to use the exact phrase.
Exact words still help in one place: the questions on your page. When a heading uses the same words a customer would use, it's easier for a platform to see that the passage under it answers their question.
In AnswerOrbit
Our counting has its own version of this problem. We count an answer as naming you when it uses your brand's name, or another name you've given us, as a whole word. So add every name and spelling your customers use in Settings, under Names. See Settings.