SEO Content Audit: What to Refresh, Merge, and Delete

An SEO content audit is a review of every page you have published, where you score each one on real search data and give it a single action: keep it, refresh it, merge it into a stronger page, or delete it. The goal is not tidiness. The goal is to stop spending time on pages that will never rank.

Most audit guides are written for sites with thousands of URLs and a paid tool budget. This one is built for a library of 20 to 100 posts and a founder with a few hours a week.

This is how to do a content audit without a tool budget. You will get the exact numbers that trigger each action, the order to work in, and the steps to merge posts without losing the rankings you already have.

What an SEO Content Audit Actually Decides

Every page in your library ends up in one of four buckets. That is the whole job.

• Keep. The page is stable or growing. Leave it alone.
• Refresh. The page has real demand but is slipping or stuck. Rewrite and update it.
• Merge. Two or more pages chase the same query. Fold them into one.
• Delete. The page has no traffic, no links, and no job to do. Remove or redirect it.

The hard part is not knowing these four options. It is deciding which one a given page gets without agonizing over it for twenty minutes.

That is why this guide leans on thresholds. A number you can check beats a gut feeling, especially when you are auditing your own writing and every post feels worth saving.

Start With a Content Inventory You Can Finish in One Sitting

Before you judge anything, you need one row per page and a few columns of honest data. This is your content audit template, and it fits on a single sheet.

Pull your URL list from your sitemap. Then open Google Search Console and export the Pages report for the last 12 months.

Here are the columns worth having. Anything more is procrastination dressed up as research.

Sort by impressions, high to low. Pages with impressions and few clicks are your best candidates, because Google already shows them to people.

If your top queries look scattered or thin, the problem may sit further upstream. A quick pass on how to do keyword research will tell you whether the topics were ever worth targeting.

The Four Actions and the Thresholds That Trigger Them

Use this table as your first pass. Work down the rows and stop at the first one that fits.

Refresh When Demand Is Real but the Page Is Stuck

Positions 11 to 30 are the sweet spot. Semrush calls these striking distance pages, and they are the cheapest wins you have. Google already thinks the page is relevant. It just does not think it is the best answer yet.

A real content refresh means new sections, current numbers, better examples, and a tighter match to what the query actually asks. Changing the date and swapping two words is not a refresh.

Open the top three results for your target query and list what they cover that you do not. That gap is your refresh brief. Work through it in this order:

• Add the missing sections your competitors all include.
• Replace every stat and price older than 18 months.
• Rewrite the opening so it answers the query in the first 60 words.
• Cut the padding you wrote to hit a word count.
• Add internal links to the posts you have published since.

That last point is easy to forget. A post from two years ago links to nothing you wrote afterward, so it sits stranded while newer pages carry the topic.

Merge When Your Own Pages Compete

Open the Queries report for two suspect URLs. If the same query shows up for both, you have split your own signals across two weaker pages.

Merging fixes this faster than any rewrite, because you are combining two partial answers into one complete one.

Delete When the Page Has No Job

Be strict here. A page with no clicks, no links, and no role in your funnel is costing you crawl attention and diluting your topic focus.

The exception is pages that serve a purpose beyond search, like a pricing page or a policy page. Those stay regardless of traffic.

Keep When the Page Is Working

If clicks are flat or rising, do nothing. Resist the urge to fiddle with a page that ranks. You can lose a good position by rewriting a page that never needed it.

How to Spot Content Decay Before It Spreads

Content decay is the slow slide a page takes after it peaks. Traffic does not fall off a cliff. It drifts down over months while you look at your site total and see nothing wrong.

The site total hides it because your new posts offset the losses. That is why you audit at the page level, not the site level.

Three signs a page is decaying:

1. Impressions hold steady but clicks fall. Your title no longer wins the click.
2. Average position slips from 4 to 9 over two quarters. Someone published a better answer.
3. The page ranked for 40 queries last year and 12 now. Its topic coverage has gone stale.

Catch a page at position 6 and a refresh can push it back up. Catch it at position 24 and you are starting over.

Google's guidance on creating helpful content is the standard your refresh should aim at. Ask whether the page still answers the question better than the pages now above it.

How to Merge Two Posts Without Losing Rankings

Merging scares people because it involves deleting a URL. Done in the right order, it is safe.

1. Pick the winner. Choose the URL with more backlinks or the better average position.
2. Move the good parts across. Pull the unique sections, examples, and data from the loser into the winner.
3. Fix the structure. The merged page needs one clear intent, not two stitched together.
4. Redirect the loser. Point it at the winner with a 301 redirect.
5. Update your internal links. Any link pointing at the old URL should now point at the winner.
6. Resubmit the winner in Search Console and wait. Give it 4 to 8 weeks before you judge the result.

Step 5 is the one people skip. A redirect that every internal link has to hop through wastes the link value you just tried to save.

If the merged page needs a full rewrite rather than a stitch job, the process in how to write SEO content covers structuring it around a single intent.

Delete, Redirect, or Noindex: Picking the Right Exit

Deleting is not one action. There are three exits and they are not interchangeable.

• 301 redirect. Use it when the page has backlinks or traffic and a close relative exists. The value transfers.
• Delete and return 410. Use it when the page is genuinely dead. No links, no traffic, no relative worth pointing at.
• Noindex. Use it when the page is useful to visitors but has no business in search results. Thank you pages are the classic case.

Never redirect a deleted page to your homepage in bulk. Google treats an irrelevant redirect as a soft 404, so you gain nothing and confuse your own reporting.

If you need to pull something out of the index quickly, Google documents the removal options and how long each takes.

Remove deleted URLs from your sitemap at the same time. A sitemap full of dead links is a slow drip of wasted crawl budget.

The Order to Work In When You Only Have a Few Hours a Week

This is where most guides leave you stranded. You end up with 60 rows and no idea what to do on Monday morning.

Work in this order. It is sorted by payoff per hour, not by what feels urgent.

1. Merge the pages that compete with each other. No new writing, fastest gain.
2. Refresh the striking distance pages at positions 11 to 30. Real demand, small gap to close.
3. Refresh the decayed pages that used to rank on page one. You are recovering traffic you already earned.
4. Clear out the dead pages. Quick, and it sharpens your topic focus.
5. Then write new posts. Not before.

That last line matters. Publishing new content on top of an unaudited library is how founders end up with 80 posts and flat traffic.

You are overdue for an audit if any of these sound familiar:

• You have published for over a year and never removed a single page.
• Your total clicks are flat while your post count keeps climbing.
• You cannot name your top five pages by traffic without checking.
• Two posts on your site cover roughly the same topic and you know it.

Three of those four are fixed by the first two steps above. None of them are fixed by publishing more.

Once the backlog is clean, feed what you learned into your plan. The patterns you spot here should shape how you build a content strategy going forward, and the mechanics in how to rank on Google apply just as much to a refresh as to a new post.

How to Tell Whether the Audit Worked

An audit is easy to fake. You can mark 60 rows, feel productive, and change nothing about your traffic.

Measure it at the page level, on the pages you actually touched. Site totals move for too many reasons to prove anything.

Track three things for each edited page, 8 weeks after the change:

• Average position, compared against the 8 weeks before.
• Clicks, compared against the same prior window.
• Number of ranking queries, which shows whether coverage widened.

Expect a mixed result. Some refreshed pages jump, some sit still, and a few dip before they recover. Two wins out of five edits is a normal outcome, not a failure.

Note the date of every change in your sheet. Without it you cannot tell a refresh effect from a seasonal swing or an algorithm update, and you will draw the wrong lesson.

Set a reminder for the review instead of trusting yourself to remember. Folding it into whatever drives your content calendar works better than a standalone task.

Content Audit Tools and What They Are Worth

You need less than the tool vendors suggest. Search Console plus a spreadsheet gets you 80 percent of the way.

A paid content audit tool earns its keep when you have hundreds of URLs, or when you want backlink data per page. Ahrefs publishes a thorough decision tree for that scale, and Semrush automates much of the data pull.

For a library under 100 posts, both are optional. Spend the money on writing time instead.

The bigger constraint is usually cadence, not tooling. Audits are worth doing precisely because they are rare, and new content still has to ship in between. Teams that struggle with both often fix the publishing side first, either by tightening their content marketing workflow or by handing the research and drafting loop to software.

That is the job SEO automation tools take on. Lunroo, for example, runs keyword research through to publishing at around 30 articles a month for $199 a month, with the first month at $99, and tracks how each page ranks afterward. It holds every draft at an approval gate, so the audit calls in this guide stay yours to make.

Frequently Asked Questions

How often should I do an SEO content audit?

Twice a year is plenty for a library under 100 posts. Quarterly makes sense once you publish weekly, because decay accumulates faster. A full audit more often than that mostly generates busywork.

Should I delete old blog posts?

Only when they have no traffic, no backlinks, and no purpose beyond search. Deleting content is not a ranking tactic on its own. The gain comes from removing pages that split your topic focus, not from lowering your page count.

Does changing the publish date help a refreshed post?

Only when you genuinely rewrote it. Google's guidance on publication dates expects the date to reflect real changes. Bumping the date on an untouched page is a trust problem waiting to happen.

What is content pruning, and is it the same as a content audit?

Content pruning is the delete and merge half of an audit. An audit covers all four actions, including keeping and refreshing. Pruning alone tends to remove pages that a refresh would have saved.

Where do I start if I have never audited my site?

Export 12 months of Search Console data and sort by impressions. Find the two pages competing for the same query and merge them. One merge teaches you more than a week of reading about audits, and the framework in SEO for startups covers the wider context.

Start With One Merge This Week

An audit is not a weekend project you complete once. It is a habit: pull the data, sort by impressions, and give every page a single clear action.

Do the merges first, refresh what sits at positions 11 to 30, and delete without sentiment. Then get back to publishing, because a clean library still needs a steady cadence to compound.

If holding that cadence is the part that keeps slipping, seeing what a research to publish loop looks like on autopilot is a reasonable next step.