A content audit inventories everything published, assesses how each piece performs, and decides what to do with it. It is unglamorous, and on a site with a few years of history it routinely finds that a third of the content is doing nothing and some of it is actively harmful.
Step 1: Build the inventory
Start with a complete list of URLs. A crawl of the site plus the XML sitemap plus a CMS export, deduplicated, catches pages that any one source misses.
For each URL, gather: title, publication date, last modified date, word count, author, content type, and whether it is indexable.
Then join performance data. From search console: clicks, impressions, average position and the top query over the last twelve months. From analytics: sessions, engagement and conversions. From a link tool, if available: referring domains. From the crawl: internal links in and out.
Twelve months of data matters, because three months will misjudge anything seasonal.
Step 2: Assign a verdict
Four possible outcomes per page. Resist adding a fifth.
Keep. Performing well, factually current, no action needed. Note the date you checked.
Improve. Has potential that is not being realised — ranking on page two, declining from a good position, or ranking for queries it answers poorly. See content refresh strategy for what this involves.
Merge. Overlaps with another page. Consolidate into the stronger one and redirect.
Remove. No traffic, no links, no strategic purpose, and no realistic path to any. Redirect to the most relevant alternative, or let it 410 if there is genuinely nothing relevant.
Step 3: Apply rules, then judgement
On 200 pages, a first pass by rule saves hours. Something like: pages with no clicks and no impressions in twelve months and no backlinks are candidates for removal; pages in positions 11–20 with meaningful impressions are candidates for improvement; pages sharing a top query with another page are candidates for merging.
Then review every candidate by hand. Rules misclassify reliably in three cases: pages that serve a commercial purpose without organic traffic, pages that are new and have not had time, and pages that support a cluster structurally even if they are not entry points.
The judgement pass is where the audit earns its value. An audit executed purely on rules will delete something important.
Step 4: Sequence the work
Order matters, because the actions interact.
- Removals first. They are quick and they clean the picture for everything else.
- Merges second. These change which URLs exist, so doing them before the improvement work avoids improving a page you are about to delete.
- Improvements third, prioritised by potential traffic rather than by how bad the page is.
- Internal links last, once the final URL set is settled. See internal linking strategy.
Executing removals without damage
Deleting content feels risky and is usually safe if done carefully.
Redirect each removed URL to the most closely related surviving page, one at a time rather than in bulk to the homepage. Bulk redirects to the homepage are treated as soft 404s and pass nothing.
Where no related page exists, returning 410 Gone is cleaner than a redirect to something irrelevant. It tells the search engine the page is intentionally gone.
Keep an archive of the removed content. Occasionally a page turns out to have had value you did not measure, and restoring it is much easier than rewriting it.
Finally, update the sitemap and check for internal links pointing at removed URLs.
What audits typically find
On a 200-page site with three or four years of history, a common distribution is roughly: 15–25% keep as is, 30–40% improve, 10–15% merge, and 25–35% remove.
The removal figure surprises people and is usually driven by three things: old company news nobody searches for, thin pages created to target keyword variations, and content about products or features that no longer exist.
What to do with the findings beyond the page list
The audit produces strategic information that gets discarded if nobody looks for it.
Which topics perform and which do not, which tells you where to invest. Which content formats work for your audience. Which authors or sources produce content that lasts. And which gaps exist — topics the cluster structure implies you should cover and do not.
That last one usually produces the next six months of the content plan, which makes the audit a planning exercise as much as a cleanup.
Keeping it current
A full audit annually is reasonable. Between audits, two lighter habits keep the inventory from drifting: review the previous quarter’s publishing against its targets, and run a monthly check on pages that declined significantly.
Maintaining the inventory spreadsheet as a living document, updated when content is published or changed, turns the next audit from a two-week project into a two-day one.
The spreadsheet columns that matter
An audit lives or dies on whether the inventory is usable. A workable column set, in order:
- URL, title, content type, author, published date, last modified date.
- Clicks, impressions, average position and top query, each over twelve months.
- Sessions, engagement rate and conversions.
- Referring domains and internal links in.
- Word count and indexability.
- Target query, as originally intended — frequently different from the query it actually ranks for, and that gap is informative.
- Verdict, reason, owner and target date.
The last four columns are what turn an analysis into a project plan. An audit that produces a verdict column with no owner and no date produces no change.
Getting agreement on removals
The technical work is straightforward; the organisational work is persuading people to delete things they wrote.
Three things help. Present the data per page rather than in aggregate — “this page had 11 visits in twelve months and no links” is hard to argue with. Frame it as concentration rather than loss: the same effort spread across 120 pages instead of 200 produces better pages. And keep an archive, so “we can restore it” is true and the decision feels reversible.
Where a page has sentimental or political weight and no performance, noindexing rather than deleting is a reasonable compromise. It removes the quality drag without the argument.
Turning the audit into a habit
The gap between a one-off audit and a maintained content programme is a small amount of recurring discipline.
Add two fields to whatever you use to track publishing: the target query, and a date for the next review. Set the review date when the piece goes live rather than deciding later — twelve months for evergreen material, six for anything that dates quickly.
Then each month, look at the pieces coming up for review and the pieces that lost the most traffic. That is a short list, it takes an hour, and it prevents the archive drifting back into the state the audit found it in.
Frequently asked questions
How long does a content audit take?
For 200 pages, roughly two to three days for data gathering and assessment, then several weeks to execute the resulting work.
Will removing pages hurt my traffic?
Removing pages that have no traffic cannot cost traffic. Done carefully with appropriate redirects, audits usually improve overall performance by concentrating signals on fewer, better pages.
Should I audit non-blog pages too?
Yes. Landing pages, product pages and help documentation all accumulate the same problems, and help content in particular is often the most neglected and the highest-traffic part of a site.
What if a page has backlinks but no traffic?
Never delete it outright. Either improve it so it earns traffic, or redirect it to a relevant page so the link value is retained.
