Search algorithm
A search algorithm is the set of rules a search engine uses to decide which pages to show for a search, and in what order.
A search algorithm is the set of rules a search engine uses to pick the pages that match a search and put them in order. When people talk about "the Google algorithm" they mean all of those rules working together, from which pages get into the index to which one ends up first.
For a marketer, the algorithm is the reason a page ranks or doesn't. Most of SEO is working out what it rewards and making pages that earn it honestly.
One name for many systems
There isn't a single formula. Google describes many separate ranking systems, each looking at something different: how well a page matches the words, how useful and original the page seems, whether its news is fresh, whether it's spam. Its ranking systems guide lists the ones it talks about in public.
The systems change all the time, mostly in small ways nobody notices. A few times a year Google makes a bigger change to its core systems, and rankings across many sites move at once.
What changes with AI search
AI platforms add a second step. A search algorithm still finds and orders pages, and then a language model reads some of them and writes the answer. So two sets of rules decide whether you're named: the search that picks the pages, and the model that picks what to say from them.
For example, Fernhill Coffee's guide to grinders might rank fifth on a search the platform runs. If the model only reads the top three results, the guide never reaches the answer. If it does read it, the model still decides whether Fernhill is worth naming.
The second step is less predictable than the first. Ask the same question twice and you can get different brands, because the model doesn't pick its words the same way every time.
In AnswerOrbit
We can't see inside any platform's rules, so we watch what comes out. Every night we ask each of your prompts on each platform and keep the answer with the pages it used, and on ChatGPT, Perplexity and Copilot, the searches it ran. How we count explains how we ask.