A search algorithm is the collection of rules and systems a search engine uses to decide which pages answer a search and in what order to show them. When people talk about “the Google algorithm”, they mean this whole machinery, not a single formula.
How a search algorithm works
Before any ranking happens, a search engine has to find and store pages. Its bots crawl the web, the pages are processed and the useful ones are added to an index. When someone searches, the algorithm works through that index in stages: it interprets what the person means, retrieves pages that could be relevant, then scores and orders them.
Google describes this as many ranking systems working together rather than one algorithm. Some interpret language, such as RankBrain, neural matching and BERT, which help match a search to pages that use different words for the same idea. Others assess the quality and usefulness of content, the reliability of a source, how fresh results need to be, whether pages load and work well on a phone, and whether a page or site is spam. The location of the searcher also feeds in, which is why someone in Leeds and someone in Bristol see different results for “accountant”.
The weightings are not published, and they shift with the query. A search about medication leans heavily on trustworthiness; a search for a football score leans on freshness. Talk of a fixed list of ranking factors with set percentages is guesswork.
Why it matters
Google changes its systems constantly, with thousands of small adjustments each year and a few larger core updates that it announces on its Search Status Dashboard. Each large change re-evaluates which pages best serve searchers, so a UK business can gain or lose visibility without having touched its site. Understanding that the algorithm rewards usefulness, not tricks, is what lets you build a site that holds up through those changes.
The same systems increasingly feed AI-generated answers. Google’s AI Overviews and AI Mode draw on pages from its index, so the work that helps a page rank in the ordinary results also affects whether it is used as a source in those answers.
Common mistakes
- Chasing individual signals. Adding keywords to hit a density target or buying links to raise a metric works against systems designed to detect exactly that.
- Believing every rumour. Industry chatter after an update often names causes that are later disproved. Google’s own documentation is a better starting point.
- Blaming the algorithm for site faults. Many traffic drops trace back to a release, a migration or a tracking change, not an update.
- Treating Google as the only engine. Bing has its own algorithm and powers other services, and some customers find businesses through it.
How to act on it
Read Google’s Search Essentials and its guide to ranking systems, and the Search Quality Rater Guidelines, which show what Google’s human testers are asked to look for. Then judge your pages against them honestly: do they answer the search better than what currently ranks, are they written by someone with real experience, and do they work well on a phone?
Keep a dated log of site changes and check it against announced updates and SERP volatility when traffic moves. If you have lost ground after an update and want an objective view of why, an SEO audit looks at technical health, content quality and links together and sets out what to change first.