First-Pass Site Diagnosis
A first-pass site diagnosis is not an eighty-page audit. It is a day or two of triage answering one question: what is on fire here. This lesson covers the order of checks, the two traps that first passes turn up most often, and how to turn findings into priorities.
The order is built around what can be confirmed: the manual actions report, the page indexing report, and the documentation on what robots.txt actually does.
What the first pass has to answer
Three questions, and nothing else:
- Does Google see the pages that matter? If not, nothing else is relevant yet.
- Is there a human decision against the site? That is a separate case with its own procedure.
- Which findings touch revenue? Priority comes from scale and proximity to money, not from how ugly the defect is.
Trap one: robots.txt does not remove a page from the index
The most common day-one finding and the most expensive misunderstanding. The documentation puts it without qualification: robots.txt tells crawlers which URLs they may access and "is used mainly to avoid overloading your site with requests; it is not a mechanism for keeping a web page out of Google".
The second half of the trap is worse. For a noindex rule to work, the page must not be blocked by robots.txt: if it is blocked, the crawler never sees the rule, and the page can still appear in search results.
Hence a state that first passes meet constantly: a section "closed from Google" by a robots.txt directive, noindex sitting inside those pages, and the pages in the results anyway. The two mechanisms cancelled each other out. Checking it takes minutes: the URL Inspection tool for one page and the site crawler for the picture at scale.
Trap two: a fall with no stated cause
If traffic has dropped, the manual actions report opens first: a decision by a person at Google is stated there outright. An empty report is a result too — it means working through the documented order, starting with seasonality and search types rather than with rewriting copy. Wins and losses finds the affected pages and queries.
The order of checks
| Question | Where to look | What counts as an alarm |
|---|---|---|
| Any manual actions | Manual actions report | Any entry: everything else waits |
| Does Google see the money pages | Page indexing report, URL inspection | Key sections not in the index |
| What is blocked, and by what | robots.txt, meta tags, response headers | A robots.txt and noindex conflict, important sections closed |
| Where canonicals point | Duplicate URL consolidation | Canonical to the home page or another page by template |
| Does the sitemap match reality | Sitemaps, sitemap monitoring | Deleted URLs, redirects and blocked pages inside it |
| Status codes across templates | Bulk check, site audit | 4xx and 5xx on categories, redirect chains |
| What broke after the last release | Wins and losses | The release date lining up with an impressions dip |
How to prioritise the findings
Three multipliers, none of them "importance by textbook":
- Scale. A template defect hits hundreds of pages; a single-page defect hits one.
- Proximity to money. A category that sells outranks the blog, even if the blog has more errors.
- Cost of the fix. A day in a template and a quarter-long migration are findings of different classes.
A finding without those three axes is not a priority, it is a line in a list. That is why a first pass ends with a decision about what comes first rather than with an inventory of errors.
Red flags
- An entry in the manual actions report.
- Money pages absent from the index while the site is live.
- A noindex on a template that covers a whole section.
- Canonicals for an entire page type pointing at one URL.
- Widespread 5xx: discussing content is pointless until they are gone.
- A migration already under way with no redirect map — that is a project type from the lesson on types of SEO project, not a defect.
The output: one page
A first pass ends in a short document: the project type, three critical findings with their scale, what happens this week, what is deferred and why, and which data is needed to test the hypotheses. The KPIs attached to it follow the rules in the lesson on goals and KPIs.
Practice
- Open the manual actions report. Write the result down even when it is empty.
- Take five URLs that matter for revenue and inspect each one: indexed or not, and what is in the way.
- Open robots.txt and find the directives that close sections. For each, say what it should have been: noindex, a password, or nothing.
- Write down three findings and rate each by scale, proximity to money and cost of the fix.
- State one decision: what gets done first, and why that one.