How much of my site does a scan actually cover?

Discovery: the crawl starts from the entry point and follows internal links, so a page nothing links to is not reached.

The robots file: paths disallowed there are not fetched, and the report says so rather than treating them as clean.

A page budget, which on a large site means a sample rather than an inventory.

Authentication: the crawl is anonymous, so everything behind a login is outside it entirely.

Rendering: links inserted by script after load may not be present in the HTML that is read, which narrows discovery further.

Each page references stylesheets, scripts, fonts and images, and fetching all of them on a heavy site exceeds any reasonable budget.

So a subset is measured and the rest is reported as unmeasured, which is what keeps the weight total honest — it is a floor rather than an exact figure.

Resources requested by script after render are outside the inventory, since the inventory is built from what the served markup references.

Public asset inspection has its own budget, which is why the coverage figure matters most there: a missed file can be a credential that stays public.

A coverage figure near the full inventory makes an absence of findings meaningful; a low one makes it a starting point.

Read the coverage figure first, before the findings. It is what decides how much the findings mean.

A clean result over a high coverage figure is evidence that the examined surface is sound.

A clean result over a low one means the sample was clean, which is a much weaker statement and frequently the honest one on a large site.

Neither is a statement about the pages that were not reached, and no amount of clean sampling becomes proof.

Which is the same reasoning as any sampled audit: the sample size is part of the result.

Make sure internal links reach the pages that matter, since discovery follows links and orphaned pages are invisible.

Check the robots file for paths disallowed by accident, which is a common cause of a section being skipped entirely.

Publish a sitemap, which gives discovery a second source beyond link-following.

Reduce the number of near-identical URLs — parameterised variants, faceted combinations — which consume budget without adding information.

Scan sections separately where a site is large enough that one pass cannot cover it.

It is not a quality measure. A site with low coverage is not worse than one with high coverage; it is larger, or harder to crawl.

It is not the same as the number of pages indexed by a search engine, which uses its own budget and its own rules.

It is not an estimate of your site's total size, since a crawl cannot know what it did not reach.

It is a statement about this scan, which is what makes it useful for reading this scan's findings.