A content audit is a page-by-page review of the content on your site that ends with a decision for every URL. You collect performance and technical data on each page, judge it against your strategy, and decide whether to keep it, improve it, consolidate it, or remove it.

It's where technical SEO and content strategy meet. The output is a complete inventory of your pages and a decision attached to each one.

The job is to evaluate the quality, relevance, and effectiveness of the content you already have. You're hunting for content gaps, pages that aren't pulling their weight, anything outdated or thin, and the openings where you could do better.

That means collecting hard data about every page, layering your own judgment on top, and then deciding what happens to each one.

  • Keep - leave the page as it is, no changes required
  • Improve - work on changes to lift the page's performance
  • Consolidate - combine multiple pages together
  • Remove - take the URL off the site because the content has no value

An audit should also tell you where your coverage is missing entirely, which feeds back into your strategy. Let's walk through the four steps it takes to run one properly.

Step 1 - Understand your content strategy

Everything you decide in an audit traces back to your strategy.

Two questions drive it. Which buying personas are you trying to reach, and how much does each one matter? And what are you actually trying to say to them?

Get clear on this before you touch a single URL. It's what lets you judge the purpose of each piece instead of guessing.

You need to be able to look at any page and say two things about it:

  • Which buyer persona the URL is targeting
  • Which buyer stage it's targeting: Awareness, Consideration or Decision

It also helps to know the site you're working on. You don't have to audit every page, and if you know the place well, you can drop entire subfolders before you start and save yourself the work later.

Take a /legal/ section. It's a necessary evil, but I've yet to meet a prospect who reads it the way they'd read your articles. Leave it out.

Step 2 - Create your content inventory

Your content inventory is a list of every URL you plan to audit. To build it properly you need an SEO crawler.

A good crawler can render and execute JavaScript, not just read the raw source. That matters because it reflects what your customers actually see, not what the server hands over before the page assembles itself.

You don't always need every page. If the site is unfamiliar, run an exploratory crawl first, purely to discover what URLs exist.

How does a crawler work?

Crawling works by programmatically downloading the HTML of a page, pulling the internal links out of it, and repeating until no new links turn up.

A simplified version looks like this:

  1. The crawler visits the start URL (typically the homepage)
  2. The HTML from the homepage is downloaded, and the links in its a href tags are extracted
  3. The URLs from each extracted link are added to the crawl queue
  4. The crawler selects a URL from the crawl queue
  5. The HTML from this URL is downloaded, and the links in its a href tags are extracted
  6. Any new URLs from the extracted links are added to the crawl queue

Steps 4-6 repeat until no new URLs are discovered. At that point URL discovery is done, and the crawler works through the rest of the queue to download the remaining HTML.

Alongside the links, the crawler is also pulling key information out of each page's HTML. That's what becomes your crawl data.

Performing a complete crawl

When you run a full crawl, don't limit it. You want it to find as much as it possibly can, and you sort out what you don't need afterwards. But if you know a site well enough, you can exclude sections up front.

Give the crawler every source you can. Connect your analytics platform so you have accurate organic traffic and engagement data tied to each URL once the crawl finishes.

Connect Google Search Console too, which brings in rankings, clicks, and impressions.

Then add every XML sitemap the site has.

The point is to capture the entire universe of URLs that exist for the site, so you end up with one long list of everything.

Fix technical issues at the template level

The crawl also surfaces your technical problems. Missing titles, broken links, slow pages, bad canonicals. It's tempting to fix them one URL at a time. Don't.

"Technical issues almost always exist at the template level," says Patrick Hathaway of Sitebulb. Find the template behind a pattern of errors, fix it once, and every page built on it improves together. That's the fastest way to level up a site technically. For the full set of technical checks, see our guide to the SEO website audit.

Filtering out unwanted pages

For a content audit you usually only care about pages that are indexable. Those are the ones search engines can actually index and surface in results.

Indexable pages are:

  • URLs that return a 200 status code
  • URLs without a canonical pointing at a different URL
  • URLs that don't carry a noindex robots directive
  • URLs that aren't disallowed in robots.txt

Now you've got a filtered list of every URL worth considering, with the quantitative data attached:

  • Title, h1 and meta description
  • Word count of the content
  • Crawl depth of the URL
  • Number of internal incoming links
  • Visit and engagement data
  • Ranking, clicks and impressions data

Export this into a spreadsheet so you can add the qualitative layers. You may also want to bring in third-party metrics, like backlink data from Majestic or Ahrefs.

The extra columns you'll want include:

  • Content type
  • Buyer persona
  • Buyer stage
  • Answer-engine readiness
  • Accessibility issues

The last two are newer, and they matter more every year.

Answer-engine readiness asks whether an AI system could pull a clean answer from the page. Does it answer its core question near the top? Are the sections self-contained enough to be quoted on their own? Is the information current?

A page can rank well. But that doesn't mean answer engines will use it. If you want to go deeper here, we've written about how to audit content for answer-engine discoverability.

Accessibility issues records whether people using assistive technology can use the page. Missing alt text, poor heading structure, low color contrast, and inaccessible PDFs all belong here. A page that shuts out part of your audience isn't doing its job, whatever its traffic says.

Filling these columns in takes manual inspection. But it won't be long before patterns in the URL paths let you make calls in bulk.

Where AI fits

On a big site, the inventory is where AI earns its place. Siteimprove.ai analyzes content depth across a whole site, flags pages missing key topic coverage, and shows which subtopics and questions competing pages cover that yours don't. That's work that would take one of us weeks by hand.

But it's also easy to over-trust. An LLM will confidently tell you a page is "comprehensive" without checking whether it answers the query someone actually typed. So don't let AI run the audit. Let it do the volume work of reading, scoring, and sorting hundreds of pages. You keep the decisions.

Step 3 - Assess every page

Add one more column: Recommendation. Using everything you've collected, judge each page and give it one of the four statuses.

Put a notes column next to it so you can spell out the reasoning.

This is the slowest step, and I don't know a way around it. But it's also the one with the biggest payoff, because the calls you make here define your page-level content strategy from this point on.

I'd add one rule before you start. Open the page. As Hathaway puts it, "The number one thing you can do wrong is to only ever look at the data and never look at the page itself." The numbers tell you where to look. The page tells you what to do.

Here's how to pick the right status for each page.

Keep

Use this when the page is performing as expected and needs no changes. The content is still topically relevant and isn't crying out for an update.

Common examples: Case studies, evergreen 'pillar' content, FAQs, recent blog posts.

Improve

This is for pages that aren't performing well enough, aren't targeted correctly, or have aged badly.

Ways to improve a page include:

  • Adjust the persona targeting
  • Adjust the keyword targeting
  • Add or update a CTA
  • Add extra content elements (e.g. video)
  • Improve internal linking
  • Rewrite sections to bring them up to date
  • Redo outdated screenshots or images
  • Restructure it so answer engines can lift a clean answer
  • Fix accessibility issues

Common examples: Low traffic pages, older blog posts, pages with poor conversion metrics, strong pages with few incoming internal links.

Consolidate

Choose this when two or more pages are doing the same job or chasing the same keywords. Any two pages built to satisfy the same intent can cannibalize one another in the search results.

Check that they really are doing the same job first. Two pages on the same topic aren't automatically cannibalizing each other. "If it's on a different stage, if it's in a different part of the journey, if it's focusing on a variant, if it's focusing on a different industry or persona, you need it," says Jeff Coyle. "It's not optional and it's definitely not cannibalization."

How much work consolidation takes depends on how similar the pages already are. If they're nearly identical, pick the stronger one and 301 redirect the weaker page into it.

If there's overlap but the weaker page still holds something valuable, move that content onto the strong page first, then redirect.

Common examples: Articles with similar themes, old blog posts, thin content pages.

Remove

This is for pages that offer no ongoing value, or where fixing them would cost more than it's worth.

The data makes removal candidates easy to spot:

  • Pages with very little content (e.g. under 150 words)
  • Pages that got no search traffic
  • Pages that barely registered any impressions
  • Pages ranking for irrelevant keywords
  • Pages not targeted at any of your personas
  • Pages with no incoming backlinks

Don't make the call on one signal alone. A well-written, genuinely strong page can pull little traffic and few links and still earn its place by showing your expertise. Kill it and you may dent your topical authority.

Removing a page usually means deleting it from your CMS so it returns a 404 (Not Found) or 410 (Gone). Do that carelessly and you create broken links, because the site still points at a page that's gone. That's bad for users and search engines alike.

Strip out every link reference to the pages you remove. Not sure where they are? Run another crawl and it'll find them.

Common examples: Location 'doorway' pages, discontinued products, duplicate content, old 'event' pages.

Step 4 - Create a plan of action

A long list of recommendations isn't a plan. The hard part is knowing what to do first.

Bigger sites hand you more opportunities and more ways to waste effort. Resource-constrained teams can't afford to guess at which fixes will land.

So prioritize. Sort that "improve" list by two things: how much authority you already hold on the topic, and how much work the fix really takes.

Topical authority is your edge. Lean into the topics where you're already strong, because that's where you're most likely to win. The effort estimate tells you the cost. Sometimes it's a quick page update. Sometimes it's a new page plus supporting content around it.

Both signals depend on your site and the content already on it. That's why your plan won't look like anyone else's.

Once you know which pages to work on, each one raises the same question. Should you update the existing page, create a new one, or both? That's its own decision, and it's worth making deliberately rather than stuffing every related topic into the page you already have.

Measure what changes

Tie every change back to results. Connect your updates to analytics and watch whether traffic, rankings, engagement, and conversions actually move.

That measurement loop is what separates an audit from a cleanup. A cleanup makes the site look tidier. But an audit tells you which changes earned their keep, so the next round is sharper than the last.

Common content audit mistakes

Four mistakes show up again and again:

  1. Focusing only on pillar content. A pillar page without supporting content around it rarely holds its ground.
  2. Ignoring intent mismatch. A page ranks for a keyword but doesn't give the person who clicks through what they came for.
  3. Creating what you think is great instead of what users seek. Your audit data should tell you what people want, not your instincts.
  4. Removing content without understanding its value. This is the most expensive one. Find the value first. Then decide.

Take Away

An audit isn't the work. The decisions are the work. Keep, improve, consolidate, remove, then act on them in the order most likely to pay off.

Build the inventory, assess every page, ship the plan, and measure what moved. Then do it again next quarter.