Is the site: operator truly reliable for measuring your website's indexation?

Why can't the site: operator display everything?

Illyes' statement addresses a persistent belief: many SEO professionals still use site:example.com to quickly assess indexation. The problem? Google only loads a subset of results, never the complete set.

Technically, displaying millions of results through the search interface makes no sense for Google — it consumes resources without delivering user value. The site: operator was never designed as an exhaustive accounting tool, but rather as a search filter among others.

What's the difference with Search Console data?

Search Console taps directly into Google's indexation databases. No filtering, no sampling — just raw numbers. The coverage report gives you the exact count of indexed pages, excluded pages, and pages with errors.

Conversely, site: applies ranking mechanisms and deduplication that can hide pages that are technically indexed but deemed irrelevant for this display. You see 500 results? That absolutely doesn't mean you only have 500 pages in the index.

Is this limitation new?

No. Google has mentioned it before, but the practice persists. Illyes' reminder aims to definitively kill the myth: stop counting site: results as a reliable metric.

The real change is that Google is now communicating more directly about these technical limitations, whereas previously it left things vague.

• The site: operator returns a sample, never all indexed pages
• Search Console is the only reliable source for knowing the real indexation status
• This limitation is not a bug — it's an architectural constraint that Google accepts
• Site: results undergo filtering, ranking, and deduplication, which skews the count

Is this statement consistent with real-world observations?

Absolutely. For years, we've observed massive gaps between site: result counts and GSC data. Sites with 10,000 pages indexed in Search Console sometimes show only 1,200 results in site: — and vice versa.

The issue is that many clients, junior agencies, and even some SEO tools continue to rely on site: for their indexation reports. This official statement validates what we already know: it's a useless metric for measuring anything.

What nuances should be applied to this claim?

Let's be honest: site: remains useful for spot-checks. Example: verifying that a new strategic page has been indexed, checking that a sensitive directory isn't accidentally indexed, spotting duplicate content by combining it with other operators.

But for global indexation reporting, it's dead. The gap between site: and GSC can reach 50 to 80% on some sites — impossible to build a strategy on sand.

[To verify]: Google doesn't explicitly clarify which criteria govern the selection of pages displayed via site:. We assume it's related to ranking, freshness, and perceived quality — but there's no official confirmation. Intentional vagueness or technical complexity? Probably both.

In which cases does this rule cause problems?

The pinch point comes with quick audits without GSC access. When you don't have Search Console rights — because the client won't grant them, because it's a competitor you're analyzing, because it's a preliminary audit — you were already losing key data. Now it's official: you're working blind.

Second problem area: third-party SEO tools that scrape site: results to estimate indexation. This statement makes those metrics completely obsolete — yet some platforms continue to display them as flagship metrics.

What should you do concretely to measure indexation?

First rule: Search Console becomes non-negotiable. If you manage a site without GSC access, you're not managing anything — you're guessing. Request read-only rights at minimum, set up a dedicated account if needed, but never work without this data again.

Second point: stop tracking site: variations in your dashboards. It measures nothing, creates false alert signals, wastes your time. Focus on the GSC coverage report: indexed pages, valid excluded pages, errors.

What errors should you avoid in tracking indexation?

Never rely on a single number. Indexation is a moving state: pages enter, exit, move to "discovered but not crawled" status. A 5-10% delta from week to week isn't necessarily alarming.

Classic mistake: panicking because site: shows fewer results than before. That means nothing. Check GSC, cross-reference with server logs if you have them, look at crawl rate and 4xx/5xx errors.

Another trap: using third-party tools that scrape site: to "estimate" indexation. This data is now officially recognized as inaccurate — stop trusting it.

How do you verify that your site is properly indexed?

• Connect your site to Google Search Console (validated property, sitemap submitted)
• Check the Coverage or Pages report (depending on GSC version) for the real count
• Identify valid excluded pages: they're technically indexable but Google chooses not to include them (weak content, duplication, canonicalization)
• Cross-reference with server logs to spot pages crawled but not indexed
• Monitor crawl errors (4xx, 5xx, robots.txt, meta tags)
• Use site: only for spot-checks (presence of a specific page, detection of indexation leaks)
• Automate tracking via the Search Console API if you manage multiple sites or large volumes