Content Pruning: When to Delete, Merge or Refresh Old Blog Posts
Not every post is worth keeping. Here is how to decide what to refresh, merge, redirect or remove, and what Google says about each.
Content pruning is deciding which old posts earn their place: refresh the useful ones, merge the doubles, redirect or remove the rest. Google says deleting content to look fresh won’t help rankings, so each cut needs a reason.
My view on clients carries over to blog posts. Not every one is worth keeping, and the ones that cost you money are the hardest to let go.
Run the numbers on a hypothetical IT services firm that started blogging in 2020. Today it has 220 posts, which sounds like momentum until somebody opens a spreadsheet: sixty walk readers through tools the firm dropped from its stack years ago, a dozen answer the same question twelve different ways, and forty haven’t pulled a single outside visitor in twelve months. Why are they still live? Nobody on the team can say.
That’s the cost of an archive nobody minds. A wrong claim buried on page 140 of the blog still speaks for the firm, and it keeps speaking while everyone’s asleep. Content pruning is the audit that makes every page justify itself. It isn’t a purge. The advice that sells deletion as a growth tactic is at odds with Google, which says in its own guidance that deleting for the sake of it won’t help.
What to Do With Each Old Post
Every old post ends up in one of five places: left alone, refreshed, merged into a stronger post, redirected, or removed. Evidence picks the place. Age doesn’t, and neither does one traffic number read in isolation. Our content audit services page covers that review for tech companies; below are the rules, the Google documentation behind them, and the order I’d work in.
Content Pruning Starts With What Google Says About It
Start with the part people skip. Google’s guidance on creating helpful, reliable, people-first content includes a list of warning signs, and one of them reads like it was written about pruning spreadsheets: removing “a lot of older content primarily because you believe it will help your search rankings overall by somehow making your site seem ‘fresh'” Google’s own answer sits in parentheses right beside it. “No, it won’t.”
The same list asks whether you’re changing the dates on pages to look fresh when the content hasn’t substantially changed, which is the other half of the same trick. Same lever, different hand.
So why prune at all? A page can be wrong, doubled up or pointless whether or not Google is watching, and accuracy, trust and the cost of keeping a page honest are all good reasons to cut it. Those are reasons. A ranking bump from a smaller site isn’t one.
Crawl budget is the other reason people sell. Google’s crawl budget guide is aimed at sites with a million or more unique pages that change weekly, or 10,000 or more that change daily, and Google calls those figures rough estimates, which puts a 200-post blog nowhere near it. Not close. If a consultant pitches pruning for crawl budget, ask which of your 200 pages is wearing Google out.
Zero Traffic Is a Question, Not a Verdict
In December 2023, Ahrefs published a study of about 14 billion pages from its Content Explorer database, and its headline is the one every pruning article quotes: 96.55% of pages get zero traffic from Google, and another 1.94% get between one and ten visits a month.
Read the limits first. By Ahrefs’ own account the sample is small next to the 340.8 billion pages in its Site Explorer index, it leans toward the quality side of the web, and the traffic figures are estimates, so a page quietly collecting a handful of long-tail visitors can still be counted as a zero.
What it does show is that zero is normal, so a post with no traffic isn’t evidence that your blog is broken. Ahrefs lists three reasons a page ends up there: nobody searches for the topic, no other site links to it, or it doesn’t match what searchers want.
A zero-traffic post gets questions, not a verdict. If nobody searches for the topic, your salespeople may still send it to prospects. That’s a job. If it carries wrong information, it hurts you whether anyone finds it or not. If another site links to it, deleting it burns something you didn’t have to build, and republishing the same words at a new URL won’t get it back, because those links still point at the old address.
Five Outcomes for Every Old Post
That leaves five outcomes for any old post. Five. That’s it.
| Outcome | When it fits | What you do |
|---|---|---|
| Leave it | Accurate, useful, and still read by buyers or your sales team | Nothing, and leave the date alone |
| Refresh it | Right topic, stale facts, tools or steps | Rewrite what changed, then update the date |
| Merge it | Two or more posts answer the same question | Fold the best material into the strongest URL and redirect the others to it |
| Redirect it | The page has links or traffic, and a close replacement exists | Permanent server-side redirect to the page that answers the same question |
| Remove it | Wrong, thin or off-topic, with nothing close to replace it | Return a 404 or 410, then clean up internal links and the sitemap |
Notice the row that isn’t there: “delete everything under 10 visits.” A threshold like that is easy to run. Easy to regret, too.
How to Decide What Gets Cut
Build one spreadsheet with one row per post, and pull the data before anybody starts arguing, because opinions about old posts run strong and evidence about them usually doesn’t exist. Each row needs six columns:
- URL and publish date.
- Clicks and impressions from Search Console, over a long date range so one slow season doesn’t mislead you.
- Backlinks to the URL, from whatever link tool you already pay for.
- Internal links pointing in.
- Whether sales, support or a partner team actually sends it to people. Ask them. Don’t guess.
- Whether it’s still true.
Then run every post through six questions in order, and stop at the first one that gives an answer. Order matters.
- Is anything in it wrong? Fix it or pull it, because a confident wrong answer is the worst page on your site.
- Does a stronger post cover the same ground? Merge them.
- Does it have links or steady impressions? Keep it, refresh it, or redirect it to its closest replacement. Deleting is off the table.
- Does a buyer, a salesperson or a partner still use it? Keep it.
- Is it thin, and written only to catch a keyword? Remove it.
- Is none of that true? Then it has no job, and taking it down costs nothing.
I’d put the same test to a client: does it pay, does it cost, does it have a job. Different invoice, same logic. A page that can’t answer is holding space that a better page, a clearer service description or a faster template could be using.
Content Pruning by Merging: Fold Competing Posts Into One
Two posts chasing one search don’t double your odds, since they end up competing with each other, and our guide to keyword cannibalization shows how to spot it. Merging is usually the fix. Not always, but usually.
Pick the survivor first. It’s usually the URL with the strongest links and impressions rather than the newest one. Fold the useful parts of the others into it, then send the losers to the survivor with a permanent server-side redirect.
Google’s documentation on redirects recommends a permanent server-side redirect when you want to change the URL shown in search results, and says 301 and 308 are the status codes for a page that has moved for good. Permanent redirects show the new target in results. It even lists the case we have here: you removed a page and want to send people somewhere new.
Then comes the cleanup nobody does. Internal links still point at the old URLs and the sitemap still lists them, so fix both. Otherwise every internal click takes a detour you built yourself.
Redirects, 404s and 410s: Removing a Post Cleanly
Once a post is going, you still have to remove it properly, and Google’s documentation settles an argument people love having.
The page on HTTP status codes says every 4xx error except 429 is treated the same way: the indexing pipeline is told the content doesn’t exist, a URL that was already indexed drops out of the index, a newly found 404 isn’t processed, and crawling of that URL gradually slows. Both 404 and 410 are on the list. Same result. The 404-versus-410 debate doesn’t change what Google does.
The bigger trap is the page that looks deleted and isn’t. If you gut a post and leave the URL answering 200, Search Console can flag a soft 404, because the page reports success while showing an empty shell or an error message. Return a real 404 or 410. Clean beats clever.
Redirect only when something close exists. A permanent redirect to a page that answers the same question carries the reader and the links with it, while a redirect to the homepage carries nobody anywhere. That part’s my judgment, not a Google rule, and I’d take a clean 404 over a lazy redirect every time.
Refreshing the Posts That Earn Their Place
A refresh means the post changes: facts get updated, steps get re-run, retired tool names come out, and every link gets clicked. Changing the date on an unchanged page is on Google’s list of warning signs, and a buyer who sees 2026 on a 2022 answer that still names a retired product spots it too.
Another hypothetical: a post on backing up Microsoft 365 that still points readers to a product the vendor has retired. Rewriting it costs less than losing what it earned, whether that’s links, rankings or a steady trickle of the right readers, so fix the steps, test each one against the current product, and keep the URL. Delete it and you start from zero. Nobody wants that.
Refresh in order of money. Posts that back a service page you sell go first, then posts that bring in links, then the rest. A post supporting a page you want cited in AI answers outweighs one that supports nothing, and our guide to AEO content strategy for MSPs covers that side.
Change the visible date when real work happened. Not before.
Where Content Pruning Goes Wrong
Deleting by traffic alone is the classic: one number, one threshold, one afternoon, and a hundred posts are gone.
Worse is cutting a post that other sites link to, with no redirect. Somebody chose to link to you, you didn’t pay for that, and you just threw it out.
Pruning to look fresh? Don’t. Google has already told you how that goes.
Then there’s the giant batch. If you cut 80 posts one weekend and traffic wobbles two weeks later, you can’t tell which cut did it, or whether a cut did it at all, since seasonality and other site changes move traffic too. Batch the work. Date every change. If traffic still moves, our diagnosis checklist for an organic traffic drop gives you the order to check things in.
And there’s the bulk pile. A firm that published hundreds of thin posts from a tool should put them in the remove pile, and Google’s spam policies explain why. They define scaled content abuse as many pages generated primarily to manipulate rankings and not to help users, no matter how they’re made, and the first example listed is generative AI used to produce many pages without adding value. We went through where that line falls in does AI content rank.
A Plan for a 200-Post Blog
Back to the hypothetical firm with 220 posts. Week one is the spreadsheet. Only the spreadsheet. Nobody knows yet what’s worth keeping, and nobody should delete a thing until they do.
After that, sort every row into the five outcomes. Refresh the posts tied to services the firm sells, merge the dozen that answer one question and redirect the losers to the winner, and remove the abandoned ones in a second batch a few weeks later so you can see what each round changed. Slow is fine.
Then fix the system, since the system is what let the archive rot. Every new post gets an owner and a review date. A post without an owner is a post nobody fixes when it goes wrong, and within a year the archive fills back up with the same problems.
Treat the archive like a client list. Everything on it earns its place, or it goes. If you want a second pair of eyes on yours, read how we approach SEO, AEO and GEO for tech companies, then book a growth strategy call.
Straight Answers on Content Pruning
Will deleting old blog posts hurt our rankings?
Only if the post earns something. Google says removing old content to seem fresh won’t lift your rankings, and a post with links or traffic takes that value with it when it goes.
Check links and impressions first. If a post has either, redirect it to a close replacement or keep it.
Should a removed post return a 404 or a 410?
Either one works. Google’s documentation treats all 4xx errors except 429 the same and removes an already indexed URL from the index.
Pick one and stick with it. Then confirm the URL really returns that status and not a 200 with an empty page, which Search Console can report as a soft 404 because the page says everything is fine while showing nothing.
Does every zero-traffic post have to go?
Wrong test, slightly. Ahrefs found 96.55% of the pages in its 2023 study got no Google traffic, so zero is the usual score, not a sentence.
Ask whether the post is wrong, whether other sites link to it, and whether your sales team uses it. Any yes means it stays or gets fixed. Zero alone doesn’t.
How many posts should we prune in one batch?
Fewer than the spreadsheet tempts you to. Google publishes no safe number, so I’d work in batches small enough to see what changed after each one.
Log the date and the URLs for every batch. When traffic moves, that log is the only way to tell which round caused it. No log, no answer.
How often should a tech company blog run an audit?
Once a year for the full pass, with a short check each quarter. That’s my rule of thumb, not a Google requirement.
A quarterly check catches the post that went wrong after a product change. The yearly pass is where you merge, redirect and remove. Keep both small.
Do we redirect every deleted post to the homepage?
Not to the homepage. A redirect should land on a page that answers the same question the old post did.
Where no close page exists, let the URL return a 404 or 410. A reader who lands on an honest error is better off than one dropped onto a page that has nothing to do with what they searched.
Find out which of your posts should stay
We look at what each page earns, what it costs, and what to fix first.
Book a Growth Strategy Call