Search Operators: The Practical Google Search Guide

Search operators matter less as a memorized list than as a way to test a visibility hypothesis. For SEO and AI-search work, the useful question is not “Which command should I use?” but “What evidence would change my next decision?” A query such as `site:example.com "AI visibility"` can suggest indexed coverage; it cannot prove that the site ranks for the topic or that an AI assistant will cite it.

This guide uses operators as evidence-gathering tools: separating a brand’s own pages from independent mentions, finding exposed documents, identifying content patterns, and creating repeatable checks for search and AI visibility. SeoVision tracks AI visibility across 9 assistants, including prompt-level visibility, citations, competitor mentions, and sentiment. That makes operator research one input into a broader visibility workflow—not a substitute for measuring how answer engines represent a brand.

How do you use search operators in Google?

Type an operator directly beside the value it modifies: `site:seovision.io search operators`, not `site: seovision.io search operators`. Google’s official search refinement guidance confirms this syntax rule.

Use the following research sequence:

1. Write the business question in plain language.
2. Decide whether the query needs inclusion, exclusion, an exact phrase, or a field restriction.
3. Add one operator to the plain query.
4. Inspect whether the results provide evidence or merely match the wording.
5. Save the query, date, location, device, and interpretation if the check will be repeated.

For example, “Does this site publish material about technical SEO?” becomes `site:example.com "technical SEO"`. “Which competitor pages discuss AI visibility?” becomes `site:competitor.com ("AI visibility" OR GEO OR AEO)`. The second query is not a measurement of competitor quality; it is a way to assemble pages for manual comparison.

What are the four types of search operators?

A practical classification links each operator to a research task rather than presenting a flat list. The categories overlap, and Google can change how consistently an operator behaves.

Google’s developer documentation lists supported operators for website debugging and monitoring, including `site:` and `-`: Google Search operators for developers. Treat both Google’s documentation and the result page as operational references, not a guarantee that every query is a complete index report.

Which search operators matter most for SEO research?

Choose the operator according to the evidence you need. `site:` examines Google’s association with a domain; quotation marks test exact brand or positioning language; `-` removes a known source of noise; `intitle:` and `inurl:` expose content patterns; and `filetype:` finds indexed documents that may sit outside the main content workflow.

How does the `site:` operator help with website checks?

Use `site:domain.com` to build a sample of pages Google associates with a domain:

• `site:example.com`
• `site:example.com/blog`
• `site:example.com "pricing"`
• `site:example.com (GEO OR AEO)`

Compare the returned URL patterns with the site’s intended information architecture. Unexpected folders, obsolete campaigns, duplicate-looking pages, or an absence of pages for a commercially important topic are useful investigation leads. They are not proof of index completeness, canonical status, or ranking performance.

A `site:` result is directional and can vary with query wording, location, personalization, and Google’s systems. Pair the observation with a full SEO audit workflow that checks crawlability, metadata, status codes, internal links, and structured data.

How do quotation marks improve search accuracy?

Exact-phrase searches are most useful when wording itself is the object of study. Try them for:

• `"AI visibility platform"`
• `"brand name" site:example.com`
• `"competitor name" -competitor.com`

These checks can reveal whether a positioning statement is repeated across a site, whether other sites describe a brand, or whether a phrase appears in a competitor’s indexed pages. They should not be treated as semantic coverage tests: a page can address the same concept without using the exact phrase. Follow with synonym and audience-language searches, especially for GEO, AEO, and AI visibility.

How do `-`, `OR`, and parentheses control intent?

Use exclusion when a recurring result type contaminates the sample, alternatives when terminology varies, and parentheses when several alternatives belong to one concept:

• `seo audit -jobs -course`
• `("AI visibility" OR GEO OR AEO) SaaS`
• `site:example.com (ChatGPT OR Claude OR Gemini)`

A query that is too broad produces a misleading competitor set; one that is too narrow can hide relevant terminology. Record excluded terms and synonyms so another researcher can reproduce the decision. Capitalize `OR` for clarity, keep the minus sign attached to the excluded term, and manually validate complex combinations rather than assuming database-style logic.

When should you use `intitle:`, `inurl:`, and `intext:`?

Use field restrictions to test how a topic is organized, not to infer its total market demand:

• `intitle:"SEO audit"` finds pages whose titles contain the phrase.
• `inurl:blog "content strategy"` finds likely blog URLs discussing the topic.
• `intext:"backlink exchange"` looks for the phrase in page text.
• `site:example.com intitle:pricing` checks for pricing-related titles.

For competitor analysis, compare the language and page formats returned by equivalent queries across several domains. A title may be rewritten in results, a page may be omitted, and URL structure may reflect legacy publishing practices. Confirm conclusions by opening the pages or using a crawler.

What does `filetype:` reveal?

`filetype:` can surface PDFs, presentations, spreadsheets, and other indexed assets:

• `filetype:pdf "SEO strategy"`
• `site:example.com filetype:pdf`
• `filetype:pptx "content strategy"`

Use the results as an exposure and governance check. Look for outdated claims, superseded brand language, lead-generation assets that no longer match the site, and confidential material that should never have been publicly indexed. Do not assume that a result’s existence means the document receives meaningful traffic.

How can search operators improve competitor and brand research?

Do not collapse publication, reputation, and visibility into one search. Each requires a different query and a different interpretation.

For AI-search visibility, run a separate prompt set across the answer engines used by the audience. Google operator results cannot establish whether ChatGPT, Claude, Gemini, or another assistant mentions or cites a brand. SeoVision’s brand tracking and AI mention monitoring guide addresses that distinction: a brand can have indexed pages yet remain absent from, weakly represented in, or negatively described by AI answers.

Can search operators diagnose content and technical SEO problems?

They can prioritize inspection, not diagnose a site by themselves. For example, `site:domain.com -inurl:desired-folder` may reveal pages outside an intended section, while `site:domain.com intitle:"untitled"` may expose weak title patterns. These are leads for a crawler, server-log review, Search Console check, or direct page inspection.

SeoVision’s audit corpus shows why brand and machine-readability checks deserve explicit attention. Among 1,476 websites audited as of September 1, 2026, 32% failed the “Brand name search ranking” check, 24% failed “Domain Rank,” and 20% failed “Structured data for AI citation.” These figures describe observed audit outcomes, not the effect of using operators. They support a practical priority: validate branded search visibility and citation readiness instead of treating a `site:` result as proof that a site is discoverable everywhere. See SeoVision’s SEO checker and website audit guide for a broader audit approach.

What search operators cannot tell you reliably

Operators cannot reveal who searched for a name, provide a complete index count, prove rankings for every user, or show why an AI assistant selected a citation. They are public observation tools, not private analytics or controlled experiments.

They also cannot establish causation. A page appearing after a query change does not prove that the change caused a ranking improvement. Location, language, personalization, device, freshness, and Google’s systems can alter results. For a meaningful comparison, repeat the same query under documented conditions and compare it with Search Console, crawling, analytics, or AI-visibility data.

What the data does not prove

SeoVision’s audit findings describe its own corpus, not every website online. The 1,476 audited sites may differ from the broader market by industry, size, platform, geography, or the reason an owner requested an audit. The reported percentages identify failures under SeoVision’s checks; they do not prove that search operators, structured data, Domain Rank, or any other single factor caused performance differences.

Likewise, one operator result is not a trend. Repeat the query under consistent conditions and compare it with independent measurement. If a result moves once, call it a fluctuation until sustained evidence supports a trend.

Is there a cheat sheet for Google search operators?

Use this as a starting point for documented checks, not as a promise of complete or stable coverage.

Can you tell who has Googled your name?

No. Search operators cannot identify individual searchers, and Google does not provide a public list of people who searched for a name. Appropriate analytics or Search Console access may show aggregate brand-query impressions, but not the identity of each searcher.

A credible brand-monitoring process measures mentions, rankings, traffic, and sentiment without claiming to expose private search behavior. For AI visibility, add assistant mentions and citations to that measurement rather than treating ordinary Google results as a proxy.

What should you do next?

1. Write five research questions tied to a decision, such as “Which competitor pages cover AI crawler access?”
2. Convert each question into a plain query, then add one operator at a time.
3. Save the final query, date, location, device, and top observations in a shared worksheet.
4. Run `site:` and exact-phrase checks for your domain, priority competitors, and important brand terms.
5. Separate findings into indexed coverage, competitor content, independent mentions, and technical clues.
6. Verify technical clues with a crawl, analytics, Search Console, or page inspection before changing the site.
7. Run the same brand and category prompts across relevant AI answer engines, because Google operator results do not measure AI visibility.
8. Fix the highest-confidence issue first, then repeat the documented checks after the change. Treat one unusual result as a fluctuation unless repeated evidence supports a trend.

How we measured

The statistics in this article come from SeoVision’s audit corpus of 1,476 websites as of September 1, 2026. The corpus records whether audited sites passed SeoVision’s “Brand name search ranking,” “Domain Rank,” and “Structured data for AI citation” checks. It is a product audit corpus rather than a random sample of all websites, so its findings should not be generalized without that limitation.