A site's own search box returns nothing useful, and you know the page exists. Scrolling the navigation is worse. Google already indexed that content — you just need to tell it to look at one domain, which takes six characters.
How to Search a Specific Website on Google
Type site: followed by the domain, with no space, then your search terms. That is the entire syntax.
site:example.com annual report returns pages on that domain mentioning both words. Drop the terms and site:example.com alone shows what Google has indexed there.
The no-space rule is not stylistic. site: example.com with a space is read as an ordinary keyword search and the filter does nothing at all. It looks like it worked, which is what makes it annoying.
You can be more specific than a domain. site:example.com/blog restricts to a path, and a full URL confirms whether one exact page is indexed.
Why the Result Count Above Your Search Is Wrong
That number is an estimate and it should never appear in a report.
Google recalculates it loosely per request, so it changes between refreshes. It commonly disagrees with Search Console by an order of magnitude in either direction. We have seen a site show 4,000 in one and 400 in the other on the same day.
What it is good for is a rough yes or no. Search a section and get zero results, and something is genuinely wrong. Get a plausible number and move to Search Console for anything precise.
The Pages report is authoritative for totals, and URL Inspection tells you why a specific page is missing rather than just that it is.
Checking Whether a Page Is Indexed
Paste the full URL after site: and see whether it comes back. Five seconds, no login.
Two things people conflate here. Indexed means Google stored the page. Ranking means Google shows it for a query. A page can be firmly indexed and still lose to every competitor, so presence in these results says nothing about visibility.
If a page is missing, the usual causes are simple: it is too new, nothing links to it, a noindex tag is present, or robots.txt is blocking the crawl. Only Search Console distinguishes between them.
Worth knowing about subfolder searches — they miss pages more often than people expect, because matching is a literal prefix. A trailing slash or a case difference produces a false negative that looks like a real one.
Combining site: With Other Search Operators
The operator becomes genuinely useful once you layer something onto it.
site:example.com intitle:pricing— pages with the word in their title tagsite:example.com -inurl:blog— everything except one sectionsite:example.com "exact phrase"— literal string match within the domainsite:example.com filetype:pdf— indexed documents only
That last pattern is the fastest way to find documents a site never linked prominently, which is the subject of narrowing results to PDFs.
One combination to avoid: the allintitle: and allinurl: family. They claim every term after them, so pairing one with site: usually means one of the two is silently dropped. Use the singular forms instead — the operator syntax reference flags which ones combine.
Searching Within a Website Across Subdomains
The bare domain generally includes subdomains, which catches out anyone auditing a main site.
Search site:example.com and results from blog, docs and shop subdomains often appear alongside the main site. Add -inurl:www and you get the opposite view — content on every host except the primary one.
That inverted search is a fast way to spot staging environments that reached the index. It is one of the checks in our site exposure audit, alongside the fix, because an indexed staging host needs a password rather than a noindex tag.
To cover two domains at once, use site:one.com OR site:two.com. Stacking two site: operators returns nothing, because Google reads it as a contradiction.
Practical Uses Beyond Finding One Page
Three habits we would recommend to anyone maintaining a site.
Spot content drift by searching a section you rarely touch. Forgotten pages surface immediately, and old campaign landing pages are a common find.
Check a competitor's structure without a crawler. A site: search shows how they organise content and roughly how much of it there is. Pair it with an on-page SEO analyzer when a particular page is worth understanding properly.
Audit URL coverage against your sitemap. Pulling every URL from a sitemap and comparing against what appears in search is a crude method, but it finds gaps quickly.
Getting More From Site Search
Start with the plain search, then add exactly one operator at a time. Most failed queries were over-specified from the first keystroke.
Remember the count is an estimate, subfolder matching is literal, and indexed does not mean visible. Those three cover nearly every confusing result we get asked about.
For queries with several operators layered together, our Google dorking tool assembles the syntax so a stray quote does not cost you ten minutes.