Content Pruning: Which Old Posts to Delete, Merge, or Leave Alone

TL;DR: Pruning is not the same as refreshing. Refreshing updates a page you’re keeping; pruning decides whether to keep it at all. Most sites with years of publishing history are carrying posts that produce nothing and dilute the site’s topical clarity. Here’s the decision tree, with numeric thresholds, and the honest caveat that pruning is a hygiene exercise rather than a growth lever.

Table of Contents

• Pruning versus refreshing
• The four-way decision tree
• Thresholds that decide the call
• What to do with genuinely old posts
• How to execute safely
• Measuring the effect

Pruning Versus Refreshing

They get lumped together and they’re different decisions made at different times.

Refreshing assumes you’re keeping the page. You update statistics, add sections, fix outdated claims, improve internal links, and republish. The URL survives.

Pruning asks whether the page should exist. Outcomes include deleting it, merging it into another page, or noindexing it.

You prune first, then refresh what survives. Doing it the other way around means spending real editorial hours improving pages you should have merged.

Why pruning matters at all: a site’s topical clarity comes from the aggregate of what it publishes. Two hundred posts where sixty are substantive and a hundred and forty are 400-word news blurbs from 2019 makes it harder for search engines — and increasingly, AI answer engines — to determine what the site is authoritative about. It also means every internal linking decision has more noise to route around.

The honest caveat: nobody should expect a traffic spike from pruning. The gains are indirect — clearer topical focus, less internal competition, less crawl waste, and a content library you can actually manage.

The Four-Way Decision Tree

For each post, in order. Stop at the first yes.

1. Does it get meaningful traffic or conversions?
→ KEEP. Refresh if dated, but don’t touch the URL. This includes posts with low traffic that convert — a 40-visit post generating two quote requests a month is a keeper.

2. Does it have external backlinks?
→ KEEP or MERGE, never delete. Links are the hardest asset to replace. If the content is weak, merge it into a stronger page and 301 the URL so the link equity transfers.

3. Does it cover a topic another page covers better?
→ MERGE. Take anything worth keeping, fold it into the stronger page, 301 the URL. This is the most common outcome on mature sites, and it’s the same procedure as keyword consolidation.

4. None of the above?
→ REMOVE. Delete and let it 404 or 410, or noindex if you have a reason to keep it reachable. Examples: 2018 news posts about an algorithm update that no longer exists, event announcements for events that happened, “we’re excited to announce” posts, thin posts on topics you no longer serve.

That’s the whole tree. Its virtue is that it protects the two things worth protecting — traffic and links — before considering anything else.

Thresholds That Decide the Call

Vague criteria produce inconsistent audits. Use numbers. These are reasonable defaults for a local service site; adjust to your scale, and pick your thresholds before you start reviewing so you’re not rationalizing per post.

Pull one export: Search Console Performance by page, 12 months. Add GA4 sessions and conversions by landing page if you have them. Then:

The 1,000-impressions-under-10-clicks row is the most valuable one. Those pages have proven demand and a fixable problem — usually position 15–30, sometimes a title that doesn’t match intent. Refresh those first. They’re the cheapest wins in any content audit.

What to Do With Genuinely Old Posts

A specific category worth its own treatment: industry news posts. Sites that published SEO or industry news for years accumulate hundreds of posts that were relevant for a week.

“Google announced X” from 2019. If X still exists and matters, fold the substance into an evergreen page and redirect. If X was rolled back or superseded, remove it.

“We’re excited to announce” posts. Company milestones, new hires, awards. These have brand value on an About or News page but rarely belong in a blog competing for search. Consider a single news archive page with brief entries, and remove the individual posts unless one has links.

Event and seasonal posts. “Our booth at the 2021 trade show.” Remove unless there’s a link.

Old case studies. Genuinely valuable — keep and refresh with current numbers. Old case studies with real client names and real results are among the most useful assets on a service business site, and they’re commonly the most neglected.

Posts written by a former provider in a different voice. Judge on the tree, not on the voice. If a post earns traffic in an inconsistent voice, refresh the voice. Don’t delete performance for tonal reasons.

How to Execute Safely

1. Export everything first. Search Console performance by page (12 months), Search Console top linked pages, GA4 landing page sessions and conversions, and a full crawl with internal link counts. This is your evidence base and your undo reference.

2. Build one spreadsheet. One row per post, columns for each signal above, and a decision column. Fill in the decision mechanically from the tree. Resist rationalizing individual posts — the whole point of thresholds is to remove sentiment from the process.

3. Get a second opinion on the removals. Have someone who knows the business scan the removal list. Occasionally a post has strategic value the data doesn’t show — the one page a big client reads before every renewal, for example.

4. Execute in batches. Twenty to fifty posts at a time, a couple of weeks apart. If something unexpected happens to overall traffic, you can attribute it.

5. Handle redirects properly. Merges get a one-hop 301 to the destination. Removals with no destination get a 404 or 410 — do not redirect unrelated content to your homepage.

6. Fix internal links. After each batch, crawl and repoint any internal links to removed or merged URLs.

7. Update the sitemap after each batch.

8. Log what you did. Date, URL, action, reason. In eighteen months somebody will ask why a page is gone, and “it had 3 clicks in 12 months and no links” is a complete answer.

Measuring the Effect

Set expectations before you start, because the raw numbers will look alarming in the wrong way.

Total indexed pages will fall. Intended.

Total impressions will fall. Also intended — you removed pages that were collecting a handful of impressions each. This is the number people panic about. It doesn’t matter.

Total clicks should hold steady or rise slightly. This is the number that matters. If clicks fall meaningfully, you removed something you shouldn’t have; check the log and restore it.

Clicks per indexed page rises. A better health metric than either raw total, and worth tracking quarterly.

Merged pages should improve over six to twelve weeks, same as any consolidation.

Give it a full quarter before judging. And be honest in reporting it: if you tell a business owner that removing 90 posts will increase traffic, you’re probably wrong. Tell them it will make the site legible, remove internal competition, and make the next year of content decisions better. That’s true, and it’s enough.

Quick Recap

• Prune first, then refresh what survives. They’re separate decisions.
• Decision tree: traffic or conversions → keep. Backlinks → keep or merge, never delete. Better page exists → merge. Otherwise → remove.
• Set numeric thresholds before reviewing so you’re not rationalizing per post.
• Pages with 1,000+ impressions and under 10 clicks are the best refresh candidates on the site.
• Export first, batch the execution, fix internal links, and log every decision.
• Expect impressions to fall and clicks to hold. Clicks per indexed page is the honest health metric.

Frequently Asked Questions

Will deleting content hurt my rankings?
Not if you follow the tree. The risk is deleting pages with links or unnoticed conversions, which is exactly what the first two branches prevent.

Is it better to noindex than delete?
Noindex keeps the page reachable for users while removing it from search — useful for pages with a non-search purpose. For genuinely obsolete content, removal is cleaner and reduces maintenance.

How often should I prune?
Annually for a site publishing regularly. After any provider change, and before any migration, because you don’t want to migrate content you were going to remove.

What about posts that rank for the wrong queries?
Refresh rather than remove. Ranking for something means you have relevance to work with; retarget the page toward the query you want.

Does pruning help with AI search visibility?
Plausibly, and for the same reason it helps traditional search: a clearer, more focused body of content makes it easier for any system to determine what you’re authoritative about. Treat it as clarity work, not as a citation tactic.

How many posts is too many?
There’s no ceiling on good content. The question is what share of your library is substantive. If most of your posts wouldn’t survive the tree, the problem isn’t quantity, it’s that quantity was the goal.