Why a page cannot be read · blocked
The site refused the request
The server answered a plain request with a refusal rather than a page. That is a decision the site made, and it is often the right one.
The site refused the request. This is a choice the site made.
§01 · what is actually happening
What is actually happening.
A request arrived carrying an identified user agent and no browser fingerprint, and the server returned 401, 403 or 429, or a 200 whose body is a challenge page rather than the site. Either way nothing readable came back.
This is usually a bot rule at a CDN or a firewall rather than a line someone wrote about this crawler specifically. Rules aimed at scrapers and credential stuffers catch every non-browser request, and an answer engine's crawler is a non-browser request.
We record a refusal only when a second independent request path reproduces it. A single refusal that one client sees and another does not is recorded as inconclusive, because a refusal names a company and one client's word is not enough to publish that.
§02 · what it costs
What it costs.
If the refusal is intended, none. Choosing not to be read by machines is a legitimate position and plenty of publishers hold it deliberately.
If it is not intended, it is the most complete version of the problem: an assistant asked about your products gets nothing at all and answers from a marketplace listing, a review aggregator or a competitor instead.
Nothing about this appears in your analytics, because the request that failed was never a session and the visit that went elsewhere never happened.
§03 · what would change it
The fix.
If the refusal was intended, nothing needs fixing. If it was not, it usually comes from a bot rule or a firewall setting applied to all non-browser traffic rather than a decision about this crawler. Check what your CDN or WAF does with a plain server request that carries no browser fingerprint.
Who does it. Whoever owns the CDN or WAF configuration. Minutes to see what the rule does, longer to decide what you want it to do.
Check your own site.
One request, one verdict, no account. If this is the finding you get back, the page above is the one that applies.
See what a machine seesThe other findings