Content Audit

A content audit is an evaluation of a site's pages: what to keep, improve, merge and remove. The main warning: around this topic there are many numeric thresholds ("under 300 words", "competitors have over 1500") that are not in Google's documentation. It says directly that it has no preferred word count and suggests asking whether the content provides substantial value compared with other pages in results. Below: how to run an audit so that decisions rest on data and on Google's documented behavior.

The basis is Google's guidance on helpful content, the page on core updates and the spam policies. The general order of an audit is in the lesson on the audit system, and the lesson on priorities helps order the findings.

What Google counts as weak content

The documented signs are not about length. Among the examples for the manual action "thin content with little or no added value" are thin affiliate pages, content from other sources (such as scraped content or low-quality guest blog posts) and doorways. In the helpful content guidance these are questions: are you summarizing what others say without added value; is the content made primarily to attract visits from search; are you producing lots of pages on many topics hoping something works; does the content leave readers feeling they must search again. More in the lessons on helpful content and content uniqueness.

So the signs in familiar checklists must be read with care:

• "Under 300 words while competitors have over 1500." This is not a criterion: a short page can fully answer the reader's task and a long one can add nothing. The question is value compared with other results;
• "0 backlinks." The absence of links to a page is not named in the documentation as a sign of weak content;
• "High bounce rate and little time on page." This is analytics behavior data: it hints where to look, but Google does not name it a sign of thin content;
• "Zero impressions in Search Console." The Performance report records data only for links that got impressions. Pages with none are simply absent from it, so you cannot see them "with a zero": you need to compare the list of pages from a crawler or a sitemap with the report. Zero impressions say a page is not shown in search, not that it is bad.

The decision matrix (practice)

What in these rows rests on the documentation:

• Merging. For duplicates Google describes redirects and rel="canonical" as strong signals and a sitemap as a weak one. See the lesson on duplicate content;
• Redirect on removal. Do not redirect many old URLs to one irrelevant page, such as the home page: it might be treated as a soft 404. The exception is when you consolidated several pages into one. See the lesson on redirects;
• Meaningful improvements. Google advises avoiding "quick fix" changes and improving content in meaningful ways, for example by rewriting or restructuring it;
• Content at scale you don't want in search. Per the spam policies, if you're hosting such content, exclude it from Search.

On deleting

The framing "we deleted 400 pages and the rest grew" is not supported by the documentation: it is a single story. Google describes deletion as a last resort, only if the content can't be salvaged. It adds that the idea of deleting entire sections often means they were created for search engines first, not people. And it warns of a common misconception: adding a lot of new content or removing a lot of older content to seem "fresh" will not help rankings; changing dates without substantial changes will not either. So deletion is a result of assessment, not its goal, and "less but better" is a sensible heuristic but not a Google principle.

How to run a content audit

1. List the pages. From a crawler, a sitemap and Search Console reports;
2. Add data: clicks and impressions by page (see the lesson on Search Console), sessions and key events from analytics. Pages that are in the list but absent from the report are candidates for a check; pages with traffic in analytics that are not in Search Console are found by the coverage gap;
3. Find decay and overlaps. Decay analysis shows articles steadily losing traffic, and the cannibalization check shows queries for which several of your addresses compete;
4. Assess by Google's questions: does the page provide original information or analysis; does it cover the topic fully; does the reader need to search again after it. Text quality analysis and entity coverage help;
5. Make a decision by the matrix and record it for each page;
6. Measure the effect honestly. Compare with similar periods and remember: it is hard to tell whether a specific change led to growth.

What is not in the documentation

• A word threshold below which a page is "thin" and a comparison of length with competitors;
• The absence of backlinks and bounce rate as signs of thin content;
• A claim that weak pages "dilute the authority" of a site and deleting them lifts the rest;
• A way to find "pages with zero impressions" right in the Performance report.

Checklist

1. The page list comes from several sources, not only the report.
2. Assessment is by value to the reader, not by word count.
3. Duplicates are merged with a redirect or canonical.
4. Removed pages return 404 or 410; redirects lead to an equivalent page.
5. A decision is recorded for each page: keep, improve, merge or remove.
6. Deleting is considered last.
7. The result is compared with similar periods.

Mistakes

• Calling any page below a threshold "thin".
• Deleting pages en masse for "authority" or "freshness".
• Redirecting all removed addresses to the home page.
• Changing dates without changing the content.
• Judging a page only by the absence of links to it.

What comes next

Next is the backlink profile audit: after assessing the pages themselves we check the links to the site.