How Caffeine works
Caffeine, which Google completed in 2010, was a change to indexing rather than to ranking. Before it, Google built its index in layers and refreshed each layer on a schedule, so a page could sit crawled but unpublished until its layer’s turn came round. Caffeine replaced that with a system which analyses small portions of the web continuously and adds them to the index as it goes.
The practical result is that crawling and indexing became a stream rather than an event. A page discovered this morning can appear in results the same day, and a page updated today does not wait for the whole index to be rebuilt around it. Google was explicit at the time that this made results fresher and the index larger, and that it was not a change to how pages were ranked.
Why Caffeine matters
It matters mostly as context for two things people still get wrong. The first is the belief that rankings shift on a monthly cycle — a leftover from the era Caffeine ended, when the index really did rebuild in batches and results really did lurch. The second is the assumption that publishing speed is irrelevant; once indexing became continuous, being early to cover something turned into a real advantage on time-sensitive topics.
It also set the pattern everything since has followed. Google’s systems now update continuously by default: core updates get announced, many other changes do not, and the idea of waiting for a scheduled refresh has gone entirely.
Common mistakes with Caffeine
The usual one is citing it as a ranking update. Caffeine changed the plumbing. It did not decide who wins a search result, and any explanation of a traffic drop that blames it is confused about what it was.
The second is assuming that continuous indexing means guaranteed indexing. Google still decides what is worth storing, and a page can be crawled and left out of the index anyway. Speed of discovery is not the same thing as an invitation.
How to act on it
Take the lesson rather than the update. Because indexing runs continuously, the things that slow you down are nearly all on your side: a page nothing links to, a sitemap that has not been regenerated since launch, a server that answers crawling slowly, or a template that hides new content behind scripts.
When something new needs to be found quickly, link to it from pages that are already crawled often, make sure it appears in the XML sitemap, and use URL Inspection in Search Console to request indexing for that single page. If pages routinely take a long time to show up, the cause is usually structural and belongs with your technical SEO rather than with your writing.