How Many Pages Does Google Have Indexed? (And Why site: Is Wrong)

How many of your pages does Google have indexed? The fastest honest answer comes from Google Search Console, under Indexing, then Pages, where Google reports the exact number of indexed URLs it holds for your property. The site: operator gives an estimate that is frequently wrong by a wide margin, sometimes by an order of magnitude, and Google has said for years that it was never meant to be a precise count.

That distinction matters more than it sounds. Plenty of SEO decisions get made on the back of a site: number that was never accurate in the first place. This guide covers the three ways to get a real indexed-page count, why they disagree with each other, and what to do when the number is lower than it should be.

Why the site: operator is not a real count

Type site:yourdomain.com into Google and you get a result count at the top of the page. It looks authoritative. It is not. Google's own search advocates have repeatedly described it as a rough estimate optimized for speed rather than accuracy, and in practice it fluctuates wildly. Run the same query twice in a day and the number can move by thousands. Run it from a different location and it changes again.

There are two specific reasons it misleads people:

• It estimates rather than counts. Google does not walk its full index to answer a site: query. It approximates, and the approximation gets worse as the site gets bigger.
• It is scoped to what the query would return. The operator is a search, not a database dump. Pages that are indexed but heavily demoted may not surface in the count you see.

Use site: for one thing and one thing only: a quick check on a single URL. Search site:yourdomain.com/exact-page-url. If the page comes back, it is indexed. If it does not, it almost certainly is not. That binary answer is reliable. The total count is not.

The three ways to get a real indexed-page count

1. Google Search Console, Indexing, Pages

This is the authoritative source, because it is Google reporting on its own index for a property you have verified. It splits every known URL into indexed and not indexed, and every not-indexed URL carries a reason: crawled but not indexed, discovered but not crawled, excluded by a noindex tag, duplicate with a different canonical, and so on. The number at the top of the indexed tab is the number to trust.

The catch is that Search Console shows you sampled example URLs per issue, not the complete list. If 1,400 pages are sitting in "crawled, currently not indexed", you get a sample and an export capped well below your real total. You know the size of the problem. You do not get the full list of which pages are in it, which is precisely the list you need to fix anything. Our guide to the page indexing report covers how to work through those buckets in the right order.

2. A bulk index checker

Take every URL you publish, check each one's index status directly, and count. This is what a bulk index checker does, and it is the only method that gives you a per-URL answer across the whole site rather than a total with a sampled explanation. It also lets you compare two lists that should match and usually do not: the URLs in your sitemap, and the URLs Google actually has.

The gap between those two lists is the most useful number in this entire exercise. It is not "how many pages are indexed", it is "which specific pages I meant to rank are missing", and that is an actionable list rather than a metric.

3. Your sitemap submission report

In Search Console, the Sitemaps report shows how many URLs you submitted in each sitemap and how many Google has discovered from it. This is the cheapest sanity check available: submit 800, see 200 indexed, and you know immediately that three quarters of what you published is not earning anything. It will not tell you which pages or why, but it tells you within ten seconds whether you have a problem worth investigating.

Why the numbers never agree

You will get three different totals from those three methods, and that is normal rather than a sign something is broken. They measure different things.

• Search Console counts URLs in Google's index for your verified property, including URLs you may not have submitted, such as parameter variants and URLs other sites link to.
• A bulk checker counts the URLs you gave it. It knows nothing about pages you forgot existed.
• The site: operator estimates, and its estimate is the one to discard when it disagrees with the other two.

If Search Console reports substantially more indexed URLs than you publish, that is worth a look. It usually means Google has indexed parameter variants, staging URLs, paginated series, or filtered category pages you never intended to expose. Those URLs compete with your real pages for crawl budget and sometimes for rankings.

What a healthy indexed-to-published ratio looks like

There is no universal target, and anyone who gives you one is guessing. What matters is the composition, not the ratio.

The question is never "what percentage of my pages are indexed". It is "are the pages I need to rank in the index, and if not, why not".

A large ecommerce site might have 12% of its URLs indexed and be in perfect health, because the other 88% are faceted filter combinations that should never be indexed. A 40-page service site with 22 pages indexed has a real problem, because there is no plausible reason for 18 of its pages to be excluded. Look at what is missing, not at the fraction.

Two ratios that genuinely signal trouble:

• Money pages missing. Any product, service, or landing page you built to convert should be indexed. If it is not, that is direct lost revenue and it should be the first thing you fix.
• The indexed count falling over time. A steady decline usually means quality-based de-indexing, a technical regression such as a bad canonical rollout, or a template change that introduced a stray noindex. Nothing in a standard stack alerts you to this, which is why sites often discover it months later.

When the count is lower than it should be

Work through it in this order, because the cheap fixes are also the most common.

First, the accidents. A noindex meta tag left over from a staging environment, a robots.txt disallow that is broader than intended, a canonical tag pointing every page in a template at the same URL. These exclude pages by mistake and they take minutes to fix. Search Console names them explicitly, so start there.

Second, discovery. Pages nothing links to are pages Google has no reason to prioritize. Export your sitemap, export every internally linked URL, and diff them. The orphans in that gap are usually the same URLs sitting in "discovered, currently not indexed". Internal links from pages that already get crawled often are the fastest fix available, and they cost nothing.

Third, quality. Pages that Google crawled and declined to index have already been judged, and resubmitting them unchanged will not change the verdict. This is where sites that generate pages from a template at scale run into a wall: if you are publishing SEO articles at scale, the pages still have to justify their own existence individually, because Google evaluates them one at a time and near-identical siblings get consolidated or dropped. Consolidate the weak ones into something stronger, or accept that they will not index and stop spending crawl budget on them.

How to track the number continuously

An indexed-page count you check once a quarter is trivia. The same number tracked continuously is an early warning system. Pages fall out of the index constantly, after a template change, a migration, a slow-server incident, or a quality reassessment on Google's side, and nothing in your analytics will tell you. Traffic drops, and three weeks later somebody works out why.

Indexing monitors coverage across every URL you publish, on Google and Bing. You get the real indexed count rather than an estimate, a plain-English reason for each page that is not indexed, an alert when a page that was indexed drops out, and automatic resubmission through official channels once the blocker is cleared. White hat only, no spam, no PBNs, and no claim that we can force Google's hand, because nobody can.