We cap the report at 25 pages. Here is what that number really buys you, measured against our own two sites.

If someone offers to review your whole website, ask the boring question first: how many pages, and which ones? It sounds like a technicality. It is actually the whole difference between a report you can use and a vague feeling you paid for. Our own two sites taught me this with two counts I took this week, not with a line from a pitch deck.

The two numbers, measured on our own sites

Every site keeps a list of its own pages — the sitemap, a plain file that spells out every address the site wants the world to know about. I counted the entries in the two sitemaps we own, through an anonymous read. Our small service site lists 8 pages. Our other site lists 864. Same owner, same two hands — one site you could read in an afternoon, the other a small library. The number that matters is not 864 by itself; it is what any finite review can possibly do with 8 set against 864.

What 25 means next to those two numbers

Our written report is capped at up to 25 agreed pages of one live site, priced once at USD 99. On the 8-page site, 25 is more than the entire site: every page fits inside a single pass, and nothing is left out by the arithmetic. On the 864-page site, 25 is a deliberate, thin slice — so the whole value moves away from how many and toward which ones, and what you are really buying is an agreed scope that you and the reviewer both wrote down before a single request was made.

That is not a trick to give you less. It is the only way a fixed price and a fixed promise can honestly exist at the same time. A review that quietly claims to cover everything either spends three weeks and a great deal of money, or it covers nothing carefully. A plain number on the front — 25, in our case — is what lets you know where the work stops and what you will actually be holding.

What those pages turn into, on our own runs

Pages are not the same thing as findings, and this trips people up, so here is the real shape from checking our own sites. When we ran the report on the service site, the scope was 24 pages, and it produced 114 requests, 1.14 megabytes moved, and 957 rows in the finished list — of which 72 needed action, 865 came back confirmed fine, and 20 were not checkable from where we stood. There were no transport failures on that pass. On a second site, a smaller scope of 12 pages opened into 502 rows, because every page gets looked at on several separate questions, not one.

Notice what those figures are and are not. They are not a grade. 72 lines needing action on a 24-page site is not a mark out of ten and it is not a letter; it is a count of specific things to look at. And the 20 we could not check is the part I am most careful about — a row that says not checkable means we genuinely could not read it from there, not that it passed. That distinction is the entire reason a scoped pass is worth more than an uncapped vague one.

Why choosing the pages is the skill

• On a site that fits inside 25, agree that up front — you can be told the whole thing in one pass, and the word everything is an honest one.
• On a bigger site, spend your care choosing which 25: the pages a customer lands on, the ones that take an order, the ones that changed last, the ones people already find in search.
• Ask who writes the list of pages. If a reviewer will not agree a scope in advance, you are buying a mood, not a report.
• Treat any review with no cap on its number as a warning sign. Uncapped coverage at a low price means thin coverage across everything.

How the requests stay gentle and honest

One more reason the cap matters: the way we read a site is deliberately read-only and light. Ordinary GET requests, the small robots file read before any page, at least a quarter of a second between hits on one host, no more than two going at once, a timeout on each request, a ceiling on redirects, a size limit per document, and no scripts executed. That is a pass designed so you never feel it in your live traffic. It is also exactly why the honest answer sometimes has to be not checkable from here — some things only appear if you run code, and running code is the one thing we will not do.

The plain lesson for whoever quotes you

Before you agree to any whole-website review, ask three boring questions and you have already done better than most: how many pages, which ones, and what happens to the ones that cannot be read. Our own two sites, counted this week at 8 and 864, are the proof that full has to be a number and not a feeling. If you would like exactly that on up to twenty-five pages of one live site — written down line by line, so you can re-check every finding yourself — that is the one-time, USD 99 report at https://svc.openyuri.com/ . Nothing in it is a promise of ranking, traffic or being found. It is a list of what we could read, what we could not, and why — and you can hand it to any other supplier and make them answer to it.