Hummingbird was a major rework of Google's core search algorithm, announced in 2013. It aimed to interpret the meaning of a whole query, including conversational and longer questions, instead of matching individual keywords one by one. Google's ranking systems guide lists it as retired: its work has been carried forward and improved by later systems such as RankBrain, neural matching and BERT.
From strings to meaning
Before Hummingbird, many results depended on whether a page contained the same words as the query. Hummingbird made more use of context, synonyms and the relationships between words, which mattered more as people began asking full questions on phones and by voice.
It arrived a year after Google launched the Knowledge Graph, and the two fit together. Search was moving from matching strings toward understanding things: people, places, products and the relationships between them.
Effects on SEO practice
Hummingbird is a large part of why exact-match keyword targeting lost value. A page could rank for queries it never used word for word, as long as it addressed the underlying need. It nudged SEO toward topic coverage and clear answers to real questions.
Hummingbird's legacy today
You will not find a separate Hummingbird update to watch for. Its role is now covered by machine learning systems that read language far more precisely. The practical advice it pointed to has not changed: write for the question a searcher is asking, not for a keyword string.
Example
One page, many phrasings
A payroll startup, Wagecrest, used to keep separate pages for “how to run payroll”, “running payroll steps” and “payroll process guide”. Since modern systems treat these as the same need, the team merges them into one thorough guide and redirects the others, which also stops the pages competing with each other.
Common mistakes
Writing one page per keyword variation
Near-identical pages for close variants split signals and often cannibalize each other.
Questions
No. It was a rework of how Google processed queries, not a system aimed at specific violations.
