How to Audit Your Domain for Google's Site Reputation Policy

How to Audit Your Domain for Site Reputation Policy
What is Site Reputation Abuse and Why Audit
Google’s spam policies define site reputation abuse as third-party content published on a host site to exploit that host’s ranking signals, usually with little or no first-party oversight and little value to the host’s own audience. The policy text is on the spam policies page, and Google’s later clarifications are posted on the Search Central blog. Your audit should be measured against that wording, not against summaries of it, including this one.
The risk sits in the parts of a domain that nobody reviews as a whole. A coupon subfolder run by an affiliate partner, a subdomain handed to an agency for “best of” roundups, an archive of guest posts that predates the current editorial team: each of these can carry the main site’s authority while following none of its content standards. The people who approved them often no longer work on the site, and no single team can list everything that lives under the hostname.
Google can respond to a violation with a manual action, which shows up in Search Console, and its spam documentation also describes algorithmic handling. In both cases you find out after the fact. A manual action report tells you a section has already been judged, and you are then cleaning up under pressure with partner contracts, redirects and stakeholders all in play. An audit run beforehand lets you decide section by section whether to keep, restructure, noindex, move or remove content while the choice is still yours.
An audit has a ceiling. A crawler can list sections, templates, outbound links and author patterns across a large site. It cannot measure intent or editorial oversight, and it cannot predict what Google will decide. Those judgments stay with a person reading the policy. The audit’s job is to hand that person a complete, evidence-backed list of places to look.
Building Your Audit Inventory
Start with a list of every hostname and path prefix that publishes under your brand. A crawler only sees what it can reach, so the inventory decides how much of the domain the audit covers.
Begin with the obvious entries: the main domain, the www variant, and each subfolder that has its own template or editorial team, such as /reviews/, /coupons/, /partners/ or /guest-posts/. Crawl the root first. Seodisias renders JavaScript, so navigation and links injected by scripts show up in the crawl and you can see which sections the main site actually points to.
Subdomains need more work, because nothing guarantees the main site links to them. Pull candidates from sources outside the crawl:
- DNS zone files and your registrar or DNS provider’s record list (A, AAAA and CNAME entries)
- Certificate transparency logs, which list hostnames that have ever received a TLS certificate
- CMS and hosting dashboards, where agencies and partners often have their own spaces
- XML sitemaps and sitemap index files, including ones referenced only in robots.txt
- Marketing, SEO and legal teams, who know about the microsites nobody documented
Feed the hostnames you collect to the crawler as separate start URLs. Each one becomes its own crawl, so a subdomain that has no inbound links from the main site still gets audited.
Know the limits before you read any results. A crawl covers the start points you supply and the links it can follow from them. Orphaned pages, sections behind a login, and hosts missing from your list will not appear, and an empty report for them means they were never visited. Write down which hostnames and prefixes you crawled and which you could not reach, because that record defines the scope of the whole audit.
Size is the other constraint. A corporate domain with hundreds of thousands of URLs can exceed the cap of a free-tier crawler (Screaming Frog’s free version stops at 500 URLs), which leaves the audit with a partial sample. Seodisias sets no URL limit and asks for no registration, and it keeps the crawl data on your machine on Windows, macOS and Linux. A complete crawl of a large domain is therefore possible without a license request, and the hostname list stays local.
Analyzing Content Structure and Link Patterns
Once you have a crawl, group the URLs by path pattern and subdomain. Put /coupons/, /reviews/, /partners/ and blog. in separate clusters, and give each cluster a row in a spreadsheet with its page count, template, average word count and dominant topic. Most sites have 8 to 15 clusters. The ones worth your time are the sections that look least like the rest of the domain.
To measure topical drift, compare each cluster against your core subject. Take the 20 most frequent terms from your main product or editorial pages, then count how many of them appear in the title and H1 of each cluster’s pages. A cluster where almost none of those terms show up is disconnected from the site’s subject. A cluster that shares the vocabulary but uses a different page template, author structure or navigation is also worth flagging. A low score alone proves nothing. A legitimate section such as a careers page will also score low. The score tells you where to read pages first.
Next, compare outbound links per cluster. Export every external link with its source URL, target domain and rel value (nofollow, sponsored, ugc or empty). Then calculate three numbers per cluster:
- External links per page. Compare each cluster against the site-wide median.
- Share of external links that carry no
relattribute. - Number of distinct target domains, and the share held by the top three.
A cluster with a high links-per-page figure, few distinct targets and no rel values behaves differently from a blog that cites many sources once each. Repeated links to the same casino, loan, supplement or coupon domains are the pattern to look for. Check the target URLs for affiliate parameters such as ref=, aff= or utm_ strings tied to a merchant.
Keep JavaScript rendering on for this crawl. Some affiliate widgets and “best of” tables inject their links after page load, so a raw HTML fetch undercounts them. Rendering is also why the URL cap matters here: a corporate domain with several hundred thousand URLs would run past the free-tier limit before the crawl covered every section. The inventory of a domain under review stays on your machine and does not go to a third-party cloud.
The numbers produce a shortlist, not a verdict. A crawler cannot tell whether a section was published to borrow the main site’s ranking signals or whether anyone on your team edited it. That judgment depends on the ownership signals you check next, and on Google’s policy text.
Detecting Ownership and Oversight Signals
With the shortlist from the link analysis in hand, turn to your sitemap and your CMS, because the signals in this section live on the page and in its configuration. Crawl each candidate section and export four things per URL: the author field, any sponsorship or partnership wording, the header and footer branding, and the indexing directives.
Authorship comes first. Check whether the section has author pages and whether those authors appear anywhere else on the domain. A coupon archive whose bylines never overlap with your editorial team’s is worth a closer look. So is a section where every article carries the same generic “Staff” label.
Next, search the rendered text for phrases such as “sponsored by”, “in partnership with”, “presented by” and “paid content”. Rendering matters again here, since many partner modules inject this wording after load. Also compare logos and footers. A section that shows a different brand name, a different contact address or its own legal entity suggests that someone outside your editorial process runs it.
Then read the technical isolation settings for each section:
- robots.txt: Note which paths are disallowed. A blocked section can’t be crawled, so your crawler will report it as empty. List it separately instead of treating it as clean.
- Canonical tags: Record where each page’s canonical points. Canonicals that lead to another domain, or to a different section, show where content actually lives.
- noindex: Count the pages with a
noindexdirective, in either the meta tag or theX-Robots-Tagheader. Confirm against Google’s current spam policies documentation how it treats noindexed content under the site reputation policy before you rely on it as a fix.
Treat everything above as a clue. A crawler reads wording and markup. It cannot tell whether your editors actually review the content, or whether a partner publishes it to borrow your rankings. Mark each section as “no signal”, “signal” or “needs a human review”, and take the flagged ones into the decision step.
Decision Framework and Conducting Your Audit
Once you have the audit data, assign each section one of four actions. The right one depends on what you find in that section, not on how risky it looks.
Keep. The section is first-party in practice: your editors commission it, review it and can answer for it. Keep it and document who owns it. If a partner contributes, add a named reviewer and a visible editorial process, and make sure the section follows the same quality bar as the rest of the site.
Noindex. The section has value for users but serves no search purpose, such as a partner archive or a coupon directory. Add noindex and leave the pages crawlable so search engines can see the directive. A robots.txt block would hide it. Check Google’s current spam policy text for how it treats this option before you rely on it.
Remove. The content has no owner, no readers and no business reason to exist. Return a 410 or redirect to the closest relevant page, and drop the URLs from your sitemap.
Move. The content is legitimate but belongs to a different business, brand or editorial standard. Host it on its own domain and redirect the old URLs with 301s. Moving a section removes its link to your site’s reputation, so involve legal and the partner before you start.
Write the decision next to each section in a spreadsheet, with the evidence behind it. Anyone who reads the sheet later should be able to see why you chose it.
Before you decide, open Search Console and check Security & Manual Actions, then Manual actions. The Search Console help center explains how to read the report and how reconsideration requests work. A notice there changes the order of work, because a flagged section needs a fix and a reconsideration request, not a slow review. An empty report does not clear you. Google can also act on this policy algorithmically, and no notice would appear.
To run the whole process in Seodisias, add every subdomain and subfolder you know as a start URL, crawl with JavaScript rendering on, and export the results. Group the export by URL pattern, then compare outbound links, link attributes and templates for each group. The crawl has no URL limit, so a 300,000-URL domain gets crawled in one pass.
Know what the crawl cannot tell you. Seodisias reaches only the pages that are linked or that you supply, so a subdomain nobody mentioned stays invisible until you find it in DNS records, the CMS or old sitemaps. It does not judge intent, it does not see who reviews content before it goes live, and it does not predict what Google will decide. It shows you which sections look different, link out heavily or carry sponsorship markers. A person then reads those sections against Google’s policy and makes the call.
To start, list your hostnames, add them as start URLs in Seodisias, and run the first crawl today.
Looking for something else?
Search every guide on the blog without going back to the list.