3-2-1 Backup Strategy: How to Never Lose Your Data
Learn what the 3-2-1 backup strategy is, how to apply it to your website, and how to confirm your backups actually work with a restore test.
View →We explain what Google dorking is, how to audit your own site with harmless search operators, and how to keep sensitive data out of search results.

What is a Google dork? A Google dork is a specialised search query built by combining Google's advanced search operators, such as site:, filetype:, intitle: and inurl:. Unlike a regular search, it narrows results to a specific site, file type or page feature, so it returns far more precise results. This article looks at Google dorking from the defensive side: why dorks matter, and what site owners should do so they do not expose their own information to search engines.
Search engines crawl and index everything on the internet that is publicly reachable. A file, a backup, a test page or a configuration file that you consider "hidden" but that sits on a server without protection will not escape the notice of crawler bots. Because dork queries make this indexed content easier to filter, they can be used by security specialists and by malicious actors alike.
So the problem is not the dork itself, but content that has been accidentally left exposed on a site. A dork only makes what is already public easier to find. That is why the right defence is to know what the search engine can see.
Most of these operators are also useful in everyday searching. All of the examples below are harmless and can be used to review your own site.
| Operator | What it does | Harmless example |
|---|---|---|
site: | Limits results to a specific domain | site:example.com |
filetype: | Lists a specific file type | site:example.com filetype:pdf |
intitle: | Searches for a word in the page title | site:example.com intitle:campaign |
inurl: | Searches for a term in the address | site:example.com inurl:blog |
"..." | Matches an exact phrase | "free estimate" |
- | Excludes a word from results | site:example.com -blog |
The key point is that the site: operator lets you quickly see how many of your pages appear on Google and which of them are indexed. If a page or file you did not expect shows up, you need to investigate it.
The first step in defence is to see what the search engine knows about you. Run the following checks on your own domain:
site:yourdomain.com and check whether the total number of results matches the number of pages you expect.site:yourdomain.com filetype:pdf to check whether documents you do not want published have been indexed. Run the same check for other types such as doc, xls, txt, zip and sql.site:yourdomain.com inurl:test or inurl:old.site:yourdomain.com -www; a subdomain you have forgotten about may still be in the index.Using this audit to look for information on someone else's site without authorisation is unethical and, in most cases, illegal. In Turkey, unauthorised access to information systems and the seizure of data are also crimes. Only carry out audits on sites you own or have written permission to test.
Database backups, environment files such as .env, configuration files, log records and source code archives should not sit in publicly accessible folders. Store these files outside the web root (the public folder) and make sure only the application can read them. Keep backups in a separate area of the server with restricted access, or move them to external storage.
If a folder has no index file and directory listing is enabled, the server displays every file in that folder as a list. That list can also be indexed by search engines. Disable this behaviour with Options -Indexes on Apache, or with the autoindex off setting on Nginx.
These two tools are often confused, which leads to serious mistakes:
noindex tag (<meta name="robots" content="noindex"> or the X-Robots-Tag: noindex HTTP header) stops the page from appearing in search results. However, the bot must be able to crawl the page to see the tag. In other words, you cannot block the same page with robots.txt and expect noindex to work.X-Robots-Tag header is used.For content that truly needs to stay private, the only correct method is authentication. noindex is not a privacy tool; it only removes content from search results.
Admin panels, staging sites and development environments should sit behind a password, an IP restriction or a VPN. Telling these environments to "keep out" with robots.txt is not enough; access itself must be restricted. Also make sure that version control folders such as .git and .svn cannot be reached through the web.
If documents uploaded by users (invoices, ID copies, contracts) sit in a public folder under predictable names, they can be indexed. Store sensitive uploads outside the web root and serve them through a script that checks authorisation.
Detailed error output should be turned off in production. Details such as file paths, database names or version numbers leak out unnecessarily. In PHP, display_errors should be off, and errors should only be written to the log file.
Shared cloud folders, publicly accessible storage buckets, code repositories and document sharing links can also be indexed. Review link-sharing settings regularly, and clean up repositories that contain secrets such as keys or passwords. Deleting a leaked key is not enough; always replace it with a new one.
noindex tag.A dork audit is not a one-off task. Repeat the same checks after every major update or migration, and whenever you open a new subdomain. For automatic warnings, keep Search Console notifications turned on, and watch your server access logs for unusual crawling activity. For a more detailed look at index management on the SEO side, our article what is SEO will also help.
Google dorking is a powerful but neutral search technique. The risk comes from data that should not be indexed being left exposed. Removing sensitive files from the web root, turning off directory listing, protecting admin and test environments, and using noindex and robots.txt correctly will largely protect your site against this kind of information leak.
If you would like to review your site's indexing and security status together, you can get in touch with us. You can find our server maintenance and security services on the products page.
Using search operators is legal, since you are performing a public search. However, gaining unauthorised access to other people's systems or using leaked data is a crime. Only carry out checks on your own site.
No. robots.txt only tells well-behaved bots your crawling preferences, and anyone can read it. Sensitive content requires authentication and storage outside the web root.
First remove the file or close access to it, then send a removal request from Search Console. If it contains passwords or keys, change them immediately.
On an Apache server, use Options -Indexes; on Nginx, use autoindex off. In most hosting panels this option is usually called directory privacy or directory browsing.
Do not use them on the same page. A bot blocked from crawling a page by robots.txt cannot see the noindex tag on it. If you want to remove a page from search results, allow crawling and add noindex.
Tell us briefly about your project, and we’ll get back to you with a roadmap and a quote.
Learn what the 3-2-1 backup strategy is, how to apply it to your website, and how to confirm your backups actually work with a restore test.
View →
Learn the essential defensive measures that protect your site, including updates, two-step verification, HTTPS, security headers, backups and monitoring, with a checklist.
View →
Learn what SPF, DKIM and DMARC records do for business email, how to set them up correctly and how to test them so your messages stop landing in spam.
View →We would like to use Google Analytics cookies to understand how the site is used. If you decline, no statistics cookies are set. Details