How Search Works and How Google Ranks
Google looks like one box, but behind it sit four separate jobs. A page can pass one and fail the next, so most SEO work is really about removing the blocker between two stages. Learn the stages and the whole field stops feeling like a bag of tricks and starts reading like a system.
The four stages
Google Search runs four stages in order. First, crawling: a bot called Googlebot finds a URL and fetches it over HTTP. Second, rendering: Google runs the page, including its JavaScript, to build the final version a browser would show. Third, indexing: Google parses that rendered page, picks a canonical version, and stores it. Fourth, serving and ranking: on a query, Google selects pages from the index and orders them.
Each stage is a gate. If Googlebot cannot fetch the page, nothing downstream happens. If the page renders empty, there is nothing to index. If the page is indexed but judged low value, it will not rank. When a page is missing from search, the useful question is not "why do I rank low" but "which of the four stages did this page fail." The mechanics of crawl, render, and index each get their own chapter. Start with crawling →
What ranking really is
Ranking is not a single score. It is many automated systems weighing hundreds of signals: how well the content matches the query, page quality, usability, links, plus the searcher's location, language, and device. There is no one factor to chase and no secret dial to turn. Content that matches intent, genuinely helps, and is technically reachable tends to win.
A few named systems are worth knowing because Google has confirmed them. RankBrain (2015) helps interpret queries it has never seen. BERT (2019) reads the relationships between words in a query, so word order and small function words change the meaning Google infers. The reviews system rewards first-hand review depth. And core updates roll out several times a year as broad changes to the overall ranking systems. Links are one of these signals, which is why they get their own study. How link authority works →
The helpful-content bar
Google used to run a standalone "Helpful Content System." In the March 2024 core update it folded that judgment into core ranking and retired the separate classifier. So this is not a penalty you recover from once. It is how core ranking judges quality continuously, more at the page level than as a single site-wide flag.
The fix is the guidance itself: write for people, not to game a ranking. Google's own self-assessment asks whether a page gives original information or analysis, whether it adds substantial value over other results, whether a reader would feel they got what they came for, and whether it shows real expertise or first-hand experience. Content quality and E-E-A-T are a chapter of their own. Content quality and E-E-A-T →
The AI Overviews shift
AI Overviews are the AI-generated summary boxes that now sit atop some results. Google rolled them out to US users by default in May 2024. They synthesize an answer from several sources and cite them, which means a searcher can get what they need without clicking any single result.
This reshapes the click, and independent studies report organic click-through drops on queries where an Overview appears (reported ranges vary widely, so treat any single percentage as directional). What you control has not changed much: rank well and give a clear, quotable structure so you are the source the answer names. That bridge from classic search to getting cited is the whole of Stage 3. What AEO is →
The floor to appear
Before any optimization, a page has to clear a floor. Google states it plainly: to be eligible for Search, a page must not be blocked from crawling, must return a successful HTTP response, and must have indexable content. Everything else in this Academy is about doing better than the floor. But if a page is invisible, check the floor first, because no amount of content or links rescues a page Google was never allowed to fetch.
A perfectly optimized site still has no guarantee of inclusion. Google says so directly. Optimization improves your odds; it does not buy a slot.
That is the mental model for the rest of Stage 1. Next, we read the goal behind a query, because matching that goal is the highest-return content decision you make. Search intent next →