How do you use a search operator?
Type the operator right before the word, with no space after the colon: site:nytimes.com works, site: nytimes.com doesn't. Stack them: site:example.com (audit OR crawl) -inurl:blog.
Google search operators and what each one does
These work in Google today; the ones that stopped are in the last section.
| Operator | What it does | Example |
|---|---|---|
"..." | Exact word or phrase (it replaced +) | "site audit checklist" |
- | Leaves out a word, site or operator | seo audit -template |
OR | Either word, in capitals | audit OR crawl |
* | Any word inside a phrase | "how to * a site audit" |
site: | One domain (subdomains included) or URL start | site:example.com |
intitle: | A word or phrase in the title | intitle:"site audit" |
allintitle: | Every following word in the title | allintitle:audit checklist |
inurl: | A word in the URL (for "cats": inurl:cats) | inurl:pricing |
allinurl: | Every following word in the URL | allinurl:seo audit |
intext: | A word in the page text | intext:"crawl budget" |
allintext: | Every following word in the text | allintext:crawl budget |
filetype: or ext: | One file type | filetype:pdf |
before: and after: | Last updated before or after a date or year | before:2020 |
inanchor: | Words in links pointing to the page; hit and miss | inanchor:"seo audit" |
Google documents only site:, filetype:, quotes, minus, before: and after:. Ahrefs still found the others working in 2023, inanchor: aside, but Google can change them without notice.
Five searches for a manual site audit
Run these on a client's domain and in a few minutes you know whether Google has what it should, and nothing it shouldn't.
What Google holds
site:example.com shows a sample of what Google holds. Ignore the count: Google says bigger sites shouldn't expect to see all their URLs.
Read the results for what shouldn't be there: old URLs, parameters, tag pages, redirects. Google hangs on to old copies.
For counts, Google's John Mueller said in 2021: "Use Search Console instead." Its Page indexing report and URL Inspection show what is indexed and why a page isn't.
MarketingTracer's SEO audit works from the other side: the pages its scan found, whether each one is indexable and, if not, why. The gap between that list and site: is your first finding.
Stray subdomains
site:example.com -site:www.example.com shows every subdomain except www: the forgotten staging site, a dev copy, an old shop. Use -site: for this. When I tried site:marketingtracer.com -inurl:www in October 2026, it gave about 213 results, all on www, so inurl: didn't look at the host name. With -site: the list came back empty, the answer you hope for.
Duplicate text
Search site:example.com "a sentence from the page": more than one result means the text sits on several pages. Swap in -site:example.com and you find copies on other sites. Duplicate content has the fixes for both.
Titles and tag pages
site:example.com intitle:"Untitled" (or "Home") finds template titles, and site:example.com inurl:tag gives a feel for how many tag pages are in Google. Neither catches every one.
Leftovers
site:example.com filetype:pdf turns up the price list from three years ago, before:2020 the pages nobody has touched since. Space these searches out: a quick run of them can bring up Google's "unusual traffic" check.
What two searches turn up on a house-clearance site
For a German house-clearance company, 78% of the URLs Google showed in 16 months brought in under 4% of its clicks. Two operator searches point straight at those pages, and the company's Search Console says what they're worth:
site:example.com inurl:tag
site:example.com intitle:"house clearance"
The first one lists tag archives. This site had 115 in Google, counting their second pages. The archives are set to be indexed and sit in the site's own sitemap. In 16 months they brought 2 clicks.
The second one, with the service word in the site's own language, brings up the district pages: 96 pages that each sell the same service in one district of a town. They're one text with the place name swapped. Of 22 compared, each shared 56% to 94% of its five-word sequences with the other pages on its template. In the last 90 days, 71 of the 85 that showed up in Google didn't get a single click.
The tag archives are the quick fix: a noindex on all of them, which in Yoast is one switch. District pages need a text of their own, or there should be fewer of them.
A search only finds what you think to type. Search Console's page list shows the URLs that got impressions, and a crawl of the whole site shows where your own links lead.
When I see a site where most of the URLs in Google bring almost nothing, I don't start writing new pages. I run two site: searches, open Search Console, and clean up first.
Which Google search operators no longer work?
+, ~, link:, info:, related: and cache: are gone, whatever older lists say.
+went in October 2011; quotes do its job.~for synonyms went in 2013.link:was officially retired in January 2017.info:stopped in 2017 and now runs a plain search.related:left Google's documentation on 18 July 2023, "as it's no longer supported".cache:followed on 24 September 2024: "The cache: search operator no longer works in Google Search."
AROUND(X) and daterange: were never documented and work only now and then; for dates, use before: and after:. A list that still calls cache: working is out of date.