Content audit: inventory every page, then keep, improve, merge or remove

What a content audit is (and is not)

A content audit answers one question for every page: is this URL doing its job for readers and for the business, and if not, what should happen to it? The output is a spreadsheet with a decision and an owner on every row, not a slide with general advice.

It is not the same as other audits you may run:

• Not a technical audit. Status codes, canonicals and indexability appear in your sheet as columns, but the question is about the content itself.
• Not only a cannibalisation check. Pages competing for one intent are one finding among many. The lesson Keyword cannibalisation covers that diagnosis in depth.
• Not a deletion project. Content pruning is a popular name for audits, and it sets the wrong expectation. Removing pages is one possible outcome, usually the smallest one.

Good moments to run one: after a sustained traffic drop, before a redesign or migration (so you do not carry dead weight to the new site), once a year on sites that publish a lot, and on any site that has published for years without a content plan.

What Google officially says

Google does not publish a content audit method, but several of its pages describe exactly the judgement an audit needs. The starting point is what its systems reward: “Google's automated ranking systems are designed to prioritize helpful, reliable information that's created to benefit people, and not content that's created to manipulate search engine rankings.”

On deleting content, the core updates guidance is the clearest statement Google has made. Read all three sentences together:

This does not say that removing pages boosts rankings. It says that improving comes first, and that removal can help only when the content was made for search engines and cannot be saved. The same page warns against cosmetic fixes: “Avoid doing "quick fix" changes (like removing some page element because you heard it was bad for SEO).” Instead, “Consider how you can improve your content in meaningful ways.”

Google's helpful content page lists warning signs that apply directly to audits and refreshes:

And on length, which many audits wrongly use as a quality score: “Are you writing to a particular word count because you've heard or read that Google has a preferred word count? (No, we don't.)”

Google also recommends the audit itself when traffic falls: “Also consider an audit of the drops you may have experienced. What pages were most impacted and for what types of searches?”

Step 1: build the inventory

You cannot audit pages you do not know about. Every source sees a different slice of your site, so combine them in one spreadsheet keyed on a normalised URL: the same protocol and host, a consistent trailing slash, lowercase, and tracking parameters such as ?utm_source= stripped. Without normalisation, https://example.com/Guide/ and https://www.example.com/guide become two rows.

A full crawl

A crawler starting at the home page finds every page reachable through links, with its status code, canonical, indexability, title and word count. It cannot find pages that nothing links to.

XML sitemaps

Sitemaps list what the CMS publishes, including pages that are not linked anywhere. They also tend to contain stale entries: redirected or deleted URLs that nobody cleaned up.

Search Console

The Performance report (Pages tab) shows the URLs Google actually displayed and how they did; Search Console keeps 16 months of data. If the interface table cuts rows off, export through the Search Console API or Looker Studio. The Page indexing report adds what is indexed and what is excluded, and why.

GA4 landing pages

The landing page report, filtered to organic search, shows sessions and key events per URL. It is the only source that tells you which pages make money or leads.

A CMS export

Posts, pages, products, tag and author archives, including old drafts and forgotten sections. It is the ground truth for what exists, and it usually shows who wrote each page and when.

Once merged, the gaps are findings in themselves. A URL with impressions in Search Console but missing from the crawl is an orphan or a legacy URL. A URL in the sitemap that returns 404 or redirects is sitemap debris. A tag archive with hundreds of indexed variants is a candidate for noindex.

Step 2: collect metrics per URL

Pull the same period for every page, and use at least twelve months if your business is seasonal, so a ski guide is not judged on its July traffic. Comparing two periods (this year and last) shows the direction of travel, which matters more than one snapshot.

Add columns for content type, topic cluster, primary query, decision, target URL (for merges and redirects), owner and due date. The decision column is the point of the whole exercise.

Step 3: judge quality with Google's questions

Numbers tell you how a page performs; they do not tell you whether it deserves to. For that, read the page. Google's self-assessment exists for exactly this: “Evaluating your own content against these questions can help you gauge if the content you're making is helpful and reliable.” The questions most useful in an audit:

• “Does the content provide original information, reporting, research, or analysis?”
• “Does the content provide a substantial, complete, or comprehensive description of the topic?”
• “Does the content provide substantial value when compared to other pages in search results?”
• “Is this content written or reviewed by an expert or enthusiast who demonstrably knows the topic well?”
• “Does the content have any easily-verified factual errors?”
• “Does your content leave readers feeling like they need to search again to get better information from other sources?”

Turn them into a written 1–5 rubric so two reviewers give the same page the same score. For example: 1 = wrong, empty or made for search engines; 2 = outdated or shallow, but the topic has demand; 3 = correct but generic; 4 = complete and trustworthy; 5 = original, first-hand, the page people bookmark. Always compare against the pages that rank for the page's main query, since a 4 in isolation can be a 2 next to the competition.

On large sites you cannot read every URL. Review pages with real demand or business value one by one, and judge long tails of similar pages (tags, old news, thin product variants) by template and by a sample. Who reviews matters too: Google suggests “consider having others you trust but who are unaffiliated with your site provide an honest assessment.” A subject expert catches errors an SEO specialist cannot. The lesson Helpful content in practice walks through the full set of questions.

Step 4: the decision framework

Two axes drive most decisions: how much value the URL holds (clicks, impressions, links, conversions) and how good, or fixable, its content is. Two special cases sit outside the grid: pages that duplicate another page's intent, and pages that visitors need but search does not.

Keep

Good traffic or conversions, and the page passes the quality questions. Fix broken links or an outdated screenshot, and leave the rest alone. Google's advice about small ranking drops applies to audits too: “There's no need to take drastic action (in fact, we recommend avoiding making changes to content that's already performing well).”

Improve

The page has demand (impressions) or business value, but the content is weak, outdated or loses to competitors. This is usually the largest group and the one with the best return. Update facts and prices, add what the ranking pages cover and yours does not, add first-hand experience, restructure for scanning, and rewrite the title if CTR is low. Google gives the same direction: “For example, it could be that rewriting or restructuring your content makes it easier for your audience to read and navigate the page.”

Merge + 301

Two or more pages answer the same intent. Pick the winner on traffic, links, conversions and intent fit (the lesson Keyword cannibalisation covers how), move the unique parts of the others into it, then redirect them. Google recommends this redirect type when a URL changes for good: “If you need to change the URL of a page as it is shown in search engine results, we recommend that you use a permanent server-side redirect whenever possible.” Update internal links to point to the winner and remove the old URLs from the sitemap.

Redirect only

The page is obsolete, but a genuinely equivalent newer page exists: last year's grant scheme and this year's, a discontinued product and its direct successor. Nothing needs merging, but the links and visitors should land somewhere useful. Google lists this case among the reasons for redirects: “You removed a page and you want to send users to a new page.” Redirect only when you are certain: “Use permanent redirects when you're sure that the redirect won't be reverted.”

Remove with 404 or 410

No meaningful traffic, no links, no conversions, no unique value, no equivalent page to send people to, and no realistic way to make it useful. Or the content is wrong or no longer true and there is nothing to replace it with. Return 404 or 410. For Google the difference is small: “All 4xx errors, except 429, are treated the same”, and “the indexing pipeline removes the URL from the index if it was previously indexed.” Many teams use 410 to record that the removal was deliberate. Then remove the URL from internal links and sitemaps. The lesson Status codes, soft 404, 410 covers the status codes themselves.

noindex

The page must exist for visitors but has no place in search: thin tag archives, internal search results, printer versions, thank-you pages. noindex keeps the page out of results but does not pass its signals anywhere, and Google can only see it if the page is not blocked in robots.txt. Never use noindex as a softer alternative to a merge: a duplicate that should be merged belongs in the merge row.

Refreshing content properly

A content refresh is an improvement you can defend line by line. If you could not explain to a reader what changed and why it helps them, it is not a refresh.

Dates: honest and consistent

Google does not trust one date signal: “Google doesn't depend on a single date factor because all factors can be prone to issues. That's why our systems look at several factors to determine our best estimate of when a page was published or significantly updated.” Help it read yours correctly:

• “Add a user-visible date to the page and feature it prominently.” Label it, for example Published or Last updated.
• In structured data, Google recommends “a subtype of CreativeWork (such as Article, BlogPosting, or VideoObject)” with datePublished and/or dateModified.
• “Ensure that the date (and optional time and timezone) match between the equivalent user-visible and structured values.”
• “Don't specify future dates, or the date of the action described on the page.”

In practice: keep datePublished at the original publication date and change dateModified and the visible date only after a substantive update. A merged page keeps the winner's original publication date.

<p>Published 12 May 2021 · Last updated 30 October 2026</p>
<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "BlogPosting",
  "headline": "How much does roof insulation cost?",
  "datePublished": "2021-05-12T09:00:00+01:00",
  "dateModified": "2026-10-30T10:00:00+00:00"
}
</script>

Process and cadence

1. Set the goal and scope. Recover from a drop, prepare a migration, or clean up the blog. Start with one section rather than the whole site.
2. Assign roles. An SEO specialist owns the sheet and the data, a subject expert scores quality, writers improve and merge, a developer handles redirects and status codes.
3. Inventory and metrics, as in steps 1 and 2.
4. Quality review with the written rubric.
5. Decisions with sign-off. Removals and merges of pages with links or conversions need a second pair of eyes.
6. Implement in batches. Merges first (they consolidate signals and remove competing pages), then improvements ranked by demand × quality gap, then redirects and removals.
7. Log every change with its date, URL, action and target.
8. Recrawl to confirm status codes, redirect targets, internal links and sitemaps.

For cadence, a full audit once a year suits most sites, with a quarterly light check of the top pages and of pages that lost traffic. Run an extra audit before a redesign or migration, and after a core update that hit you, once the rollout has finished: Google recommends “waiting at least a full week after a core update completes before analyzing your site in Search Console.”

Tracking results

Save a before snapshot of every page you touch, add a Search Console annotation or a change-log entry on each batch date, and then measure:

• Improved pages: clicks, impressions and CTR, compared over equal periods with the same weekdays, and year on year for seasonal topics.
• Merge targets: whether they picked up the queries and clicks of the pages merged into them.
• The section as a whole: total clicks and key events for the audited section, not the whole site, where other changes blur the result.
• Removed and redirected URLs: URL Inspection and the Page indexing report, to confirm Google has seen the redirect or the 404/410.

Be patient. According to Google, “some changes can take effect in a few days, but it could take several months for our systems to learn and confirm that the site as a whole is now producing helpful, reliable, people-first content in the long term.” And: “Keep in mind that there's no guarantee that changes you make to your website will result in noticeable impact in search results”.

Common mistakes

• ✕Deleting pages with links or trafficTheir links, visitors and conversions disappear; improve, merge or redirect instead.
• ✕Mass redirects to the home pageVisitors land on an unrelated page and Google usually treats such redirects as soft 404s. Redirect to an equivalent page or return 404/410.
• ✕Pruning in the belief that it boosts rankingsGoogle says removing lots of content to seem fresh won't help; removal is the last resort for content that can't be saved.
• ✕Changing dates without changing contentGoogle lists it as a warning sign; update the date only after a substantive update.
• ✕Judging by word countGoogle has no preferred word count. Short and complete beats long and repetitive.
• ✕Deciding on one metricA page with no clicks can still earn links or leads. Look at all columns.
• ✕A short data windowThree summer months will condemn every winter page. Use at least a year for seasonal sites.
• ✕Forgetting internal links and sitemapsMenus and sitemaps keep pointing at removed or redirected URLs.
• ✕noindex plus a robots.txt blockIf the page is blocked, Google never sees the noindex.
• ✕Rewriting top performersPages that already work need monitoring, not a makeover.
• ✕No change logWithout dates and actions you cannot tell which decision helped or hurt.

FAQ

Sources

• Google Search Central — Creating helpful, reliable, people-first content
• Google Search Central — Google Search's core updates and your website
• Google Search Central — Influence your byline dates in Google Search
• Google Search Central — Redirects and Google Search
• Google Search Central — How HTTP status codes, and network and DNS errors affect Google Search

Quotes checked against the official pages on 30 October 2026.