Most SEO content audits end as a spreadsheet. The spreadsheet is accurate and the priorities are defensible. The rewriting is where the budget runs out.
That gap is the operational problem. An agency can identify fifty thin pages before lunch, but replacing them takes a writer, a brief, and a round of edits and client approval for every one. Eventually the spreadsheet gets filed and the engagement moves on.
What the scan actually returns
When you point Ghosts at a client URL, it reads the full site and classifies every page by type: home page, service pages, city or service-area pages, and supporting content. Classification matters because the work to fix a thin city page looks nothing like the work to fix a thin home page, which most audit tools lump together anyway.
Beyond the existing pages, the scan surfaces gaps. A roofing contractor covering twelve counties might have service pages for four. A med spa might carry a page for Botox and nothing for filler, even though both terms drive comparable search volume in their market. Those missing pages are the ones the client often does not know they need, and they are frequently the faster wins because there is no existing page to redirect or rewrite.
What a measured band is and why it beats a word-count rule
The phrase "thin content" is not a specification. Telling a writer to add three hundred words gets you three hundred words of padding. Ghosts uses bands measured from the pages that currently rank, held separately by page type. Each band describes what pages ranking for a given query in a given industry actually look like: structural patterns, topical coverage, internal linking, and the presence or absence of specific content elements.
Those bands describe what exists at the top of results. They do not cause rankings, since links and domain authority factor in too. What a band gives you is a concrete target derived from observed data, well past the word-count heuristic someone published in 2019 and never updated.
What does a ranking page actually contain?
That is the question the bands answer. A ranking home page in a service-area business category looks different from a ranking city page for the same business. Page length varies by type and industry, but the structural signals matter more than raw word count: whether the page addresses the geographic service area explicitly, whether it covers adjacent services or only the single term it was built around, how it connects to other pages on the site.
Service pages that rank tend to be more specific about process and outcome than service pages that do not. City pages that rank include more than a paragraph of swapped-in location text. These findings are unsurprising, but they are specific enough to brief against, which a general "add more content" instruction never is.
The two gaps almost nobody fills
The typical ranking page cites no authority at all, not a study or a piece of government data or a trade publication, and almost none name a real author with a linked profile either.
Both gaps are cheap to close relative to every other content investment. A page that cites a relevant source from the CDC, a state licensing board, or an industry association costs a few minutes to update, and a named author with a linked bio and a real credential costs an afternoon to set up across a site. Whether those additions directly influence rankings is an open question, but they demonstrably affect how a reader evaluates the page, and they are signals almost no competitor has bothered with, which is why we built citation checks into the drafting step. Leave them out and you produce pages that match the exact gaps we were trying to close.
How Ghosts runs the audit and drafts the output
The process is a single system. You submit a client URL. Ghosts classifies the site, identifies thin pages and content gaps, then drafts replacements written to the measured bands for that page type and industry, with citation checks running against the draft before it surfaces. A human reviewer on your team sees the draft, approves or edits it, and it clears into the CMS.
Approval sits between the draft and the client's site by default. Ghosts runs on your behalf, and you are still accountable for what goes live, so the draft waits for a decision rather than assuming one. Holding output behind a human read is also what lets you answer a client who asks what your review process looks like, which is why most teams keep it that way even where they could hand the last step to a schedule.
The output is a page draft your team can copy and paste into whatever CMS the client runs. Audit, gap analysis, and the approval queue run as one system because the gap between an audit and a published page is exactly where agencies lose time and margin.
Run a free scan on any client URL at ghosts.app, no card required.