How Search Works: The Bedrock of Digital Visibility

Most service business owners I talk to treat search engine optimization like buying a billboard. They see it as a channel: a place where you spend money and leads eventually come out the other end.

That misunderstanding is expensive, and it's the reason I see so many companies burn through a marketing budget with nothing to show for it. Google itself rules out the billboard model. It doesn't accept payment to crawl a site more often or to rank it higher, and it doesn't guarantee that any page will be crawled, indexed, or shown, even a page that follows every rule.

Search isn't a channel. It's a system: a dynamic set of interconnected signals, entities, and relationships. You don't buy organic visibility. You engineer your business so the system can find it, understand it, and trust it.

If you want to win in search today, and lay the groundwork for however discovery works in five years, you need to understand the mechanics underneath. Without that, you're throwing things at a wall to see what sticks.

Here's how Google Search actually works, how it got this way, and what it means for your business. Other search engines run on the same principles. I use Google because it publishes the most about how its own system works, so every claim here can be checked against the source. It's the starting point for my SEO guides.

The three stages every page passes through

Google Search works in three stages: crawling, indexing, and serving search results. Not every page makes it through all three.

Google's stated mission is to organize the world's information and make it universally accessible and useful. To do that across hundreds of billions of pages, it runs every page through the same pipeline.

I want to be blunt about the order here, because it's where most SEO advice goes wrong. Almost everything sold as "SEO" is stage three work. If you fail stage one or two, stage three can't happen at all.

Crawling: how Google finds your pages

Crawling is how Google finds out which pages exist and what's on them. There's no central registry of every page on the web, so Google has to constantly look for new and updated pages.

It discovers pages three ways. It revisits pages it already knows. It follows links from known pages to new ones, including links inside your own site, such as a services page linking to a new service area page. And it reads sitemaps: files you submit to Google listing the pages you want crawled.

The software that does the fetching is Googlebot, Google's web crawler. Googlebot decides algorithmically which sites to crawl, how often, and how many pages to fetch from each. It also slows down when a server shows signs of strain, so a site that throws errors gets crawled less. Discovering a page doesn't guarantee a crawl. Pages the site owner has blocked, or pages locked behind a login, don't get crawled at all.

During the crawl, Google renders the page the way a browser does, running its JavaScript in a recent version of Chrome. That matters because many modern sites load content with JavaScript, and content Google can't render is content it can't see.

Indexing: how Google understands your pages

Indexing is the stage where Google works out what a page is actually about. Finding a page isn't the same as understanding it.

Google processes the text on the page and its key tags, such as the title element and image alt text, along with the images and video themselves. It also records signals about the page, including its language, the country its content is local to, and how usable it is. The results are stored in the Google index, a massive database spread across thousands of computers.

Two things about this stage surprise most business owners.

Google keeps one version of similar pages

Before storing anything, Google groups pages with similar content into clusters and selects the most representative one as the canonical: the version eligible to appear in results. The rest become alternates, shown only in specific contexts.

This is the indexing problem multi-location businesses actually have. Ten location pages built from the same template, with only the city name swapped, can look like duplicates. Google may keep one and set the others aside, and the city you most needed to rank in can be the one it set aside.

Indexing isn't guaranteed

Not every page Google processes gets indexed. Google names the common causes: low-quality content, a robots meta tag telling it not to index the page, and site designs that make indexing difficult.

Two related details matter here. First, robots.txt and the robots meta tag do different jobs. Robots.txt controls whether Google can fetch a page at all. The meta tag controls whether a fetched page gets indexed. Block a page in robots.txt and Google never reads its noindex tag. Second, Google uses the mobile version of your site for indexing. If your mobile layout drops content, that content drops out of Google's view.

Serving results: how Google decides what to show

Serving is the stage where Google answers a search, and ranking happens inside it. When someone searches "emergency HVAC repair near me," Google searches its index for matching pages and orders them by what it judges most relevant and highest quality.

Relevance is determined by hundreds of factors, and some of them belong to the searcher, not to you: their location, their language, and their device. The same search returns different results in Boise than in Boston.

The query also changes the shape of the results page. A search with local intent usually brings up a map with business listings. A search for something visual brings up images. A complex question may bring up an AI Overview. Which features appear depends on what Google thinks the searcher needs.

For local searches specifically, Google says its local results are based mainly on three factors: relevance (how well a business matches the search), distance (how far it is from the searcher or the location in the search), and prominence (how well known and well regarded it is, including its reviews and what the rest of the web says about it). For a service business, that map result is often where the job is won or lost.

If Google Search Console, Google's free reporting tool for site owners, shows a page as indexed but you never see it in results, Google lists three likely reasons: the content doesn't match what people are searching for, its quality is low, or a robots meta rule is preventing it from being shown.

The link graph: why authority still governs search

Links are one of the main ways Google understands both what a page is about and how much to trust it. To see why, it helps to know where Google came from.

Early search engines leaned heavily on what was written on the page itself. Repeat "plumber" often enough and you had a decent shot at ranking for plumber. The system rewarded spam, and it showed.

In 1998, Larry Page and Sergey Brin published the research behind PageRank, which borrowed from academic citation analysis. In academia, the most important papers are the ones cited most often by other important papers. PageRank applied that logic to the web: a link counted as a vote, and a vote from an important page counted for more than a vote from an obscure one.

PageRank has changed a great deal since then, but Google says it remains part of its core ranking systems, alongside other link analysis systems that use links to understand what pages are about. The underlying idea, that links reflect real-world trust and relevance, has changed less than people assume.

From keywords to context: how Google learned to read intent

Modern Google matches meaning as well as words, and that shift happened through a series of named systems over more than a decade. It's also where the gap between amateur and serious practitioners is widest.

Early search was lexical: it matched the words you typed against the words on the page. Search "emergency plumber" and the engine looked for pages containing those terms.

That matching still happens. What changed is the layer of understanding Google built on top of it:

• Hummingbird (2013) was a major overhaul of Google's overall ranking systems. Google now lists it as retired, its work carried forward by the systems that followed.
• RankBrain (2015) helps Google understand how words relate to concepts, so it can return a relevant page even when the page doesn't contain every word in the search.
• Neural matching (2018) helps Google connect the concepts in a query to the concepts on a page.
• BERT (2019) helps Google understand how combinations of words express different meanings and intent.
• Passage ranking (2021) lets Google identify a single relevant section of a page rather than judging only the page as a whole.

Someone searching "water coming through my ceiling" needs a plumber, a roofer, or a water damage restoration company, and Google can work that out even though none of those words appear in the search. Intent is the signal. Phrasing is not.

Experience, expertise, and trust

Google describes the quality it looks for as E-E-A-T: experience, expertise, authoritativeness, and trustworthiness, with trust the most important. E-E-A-T itself isn't a specific ranking factor. Google's systems use a mix of signals that tend to identify content with those qualities, and they weigh them more heavily on topics that can affect people's health, finances, or safety.

Your business is an entity

An entity is a distinct, recognizable thing with relationships to other things: a person, a place, an organization. Google maintains a database of entities and the facts connecting them, called the Knowledge Graph, and your business can be one of them, connected to an address, an owner, a set of services, and a reputation.

In my experience, the more clearly and consistently that entity is defined everywhere it appears (your website, your Business Profile, directories, review sites), the more confidently Google surfaces it. Conflicting names, addresses, or phone numbers don't just fail to help. They create ambiguity about which business is which, and ambiguity costs you placement. Google's own guidance on AI features points the same direction: it tells site owners to keep their Business Profile information current.

What this actually means for your business

Google is built to connect each search with the most reliable answer available, and every stage of that process rewards the same things: pages it can reach, pages it can understand, and a business it can trust.

If your presence is built on shortcuts, or on a misreading of how the system works, your visibility will always be fragile. It'll move when the algorithm moves. When you align with the mechanics instead, through solid technical foundations, verifiable authority from real links, and content clear enough for a machine to understand, you get visibility that compounds and that competitors can't easily take from you.

This is the bedrock, and AI search sits directly on top of it. Google states that there are no additional requirements or special optimizations needed to appear in AI Overviews or AI Mode. To be shown as a supporting link, a page has to be indexed and eligible to appear in Search with a snippet. No special files, no special schema. The foundation is the AI strategy.

The same logic carries into map pack visibility, other AI assistants, and the agentic search layer arriving now. Each one depends on whether machines can find, understand, and trust your business. Build the foundation properly and you're not just winning today's searches. You're positioned for whatever the web turns into next.