When to kill a blog post (content pruning for SEO)

Of all the maintenance jobs on a business blog, deleting posts is the one that feels most wrong. You paid for that content, in money or in evenings. It's finished. It's not hurting anyone, surely. And yet "content pruning" keeps appearing in SEO advice, sometimes with near-magical claims attached: delete your weak posts and watch the survivors soar.

The truth is duller and more useful. Pruning is a real practice with a real, modest rationale, it's occasionally the right call and frequently oversold, and the important skill isn't deleting bravely, it's telling the difference between a post that should die and a post that should be fixed. Most posts that look dead should be fixed. This article gives you the test.

Why dead pages can cost you anything at all

A page that attracts nothing seems free to keep, so start with why it might not be.

The strongest documented reason is Google's own design. The helpful content system, announced in 2022, was introduced explicitly as a site-wide signal: enough unhelpful content, in Google's words content made for search engines rather than people, can affect how the entire site is evaluated, not just the offending pages. Since March 2024 that assessment runs continuously inside Google's core systems, the same machinery that demoted mass-produced thin content across the web. A site that's one-third abandoned, outdated, or valueless pages is showing Google a worse average self than its good pages deserve.

The quieter costs are human. Every published page is a possible landing page, and a customer who arrives on a stale post with 2022 prices and dead links has met your business at its worst. Every page is also maintenance surface: one more thing to keep accurate, one more row in every audit you run.

What pruning is *not*, despite the folklore: a growth hack. Deleting pages does not, by itself, boost the survivors, and the dramatic before/after charts in pruning case studies usually belong to huge sites with tens of thousands of junk pages, plus a simultaneous quality push that did most of the work. For a small business with sixty posts, prune because dead weight has real if modest costs, and because concentration beats sprawl, not because deletion is fertilizer.

The six-part death test

A post earns deletion only by failing all six. One "yes" anywhere and it exits a different door.

• Is it older than a year? Younger posts are still in the slow ramp every post goes through. Pruning a six-month-old post is diagnosing a seed as barren.
• Does Search Console show demand? Not just clicks, impressions. A page with zero clicks but growing impressions is being auditioned by Google, and needs a fix, not a funeral. Dead means flatlined on both, for months.
• Has it converted anyone? Check key events by landing page over the full year. A post that produced two customers at zero traffic is a hidden converter with a distribution problem, the opposite of prunable.
• Does anything link to it? External links pointing at the page are accumulated authority. You don't delete pages with real backlinks; you update them or redirect them somewhere that inherits the value.
• Does it hold a structural job? Some pages exist to complete a topic cluster, catch a seasonal query, answer the question your sales conversations always reach, or serve as the reference you link from other posts. Zero-traffic pages can still be load-bearing.
• Would you write this topic today? The strategy question. If the topic was always aimed at the wrong audience, or the business has moved, no refresh can save it, and that's precisely when deletion beats optimism.

Run the test during a content audit with the data already open, and the verdicts take a minute per page. On a typical small-business blog, expect the genuine kill list to be short, a handful of pages, not a purge. If your test is condemning a third of the site, either the content program had a strategy problem worth understanding first, or your criteria are too bloodthirsty.

The three exits

Posts that fail some of the test but not all of it leave through better doors.

Update, the default exit. Proven demand plus decayed delivery equals refresh: new facts, current year, deeper answers, your experience added where the page is generic. This is the right call for the large majority of underperformers, because the demand already exists and ranking history is easier to restore than to build.

Merge, for overlaps. When two or three of your posts chase the same query and split the signals, consolidate: keep the strongest URL, fold in the unique value from the others, and permanently redirect the losers to the winner. This is pruning's genuinely high-yield form, since it converts internal competition into one stronger contender, and it preserves everything the dead URLs had earned.

Remove, for the true dead. Failed all six: delete the page and put a permanent redirect (a 301) from the old URL to the most closely related living page, its topic cluster hub, the parent category, or the closest sibling post. Redirect to the homepage only when nothing related exists at all, and accept that a pile of homepage redirects is barely better than none. If genuinely nothing relates, letting the URL return a clean "gone" status is honest and fine. What you never do is leave a farm of deleted URLs serving errors while other pages still link to them, so finish every removal session by fixing the internal links that pointed at the deceased.

The edge cases the test doesn't settle

A few page types resist the six-part test, and they're worth naming because they're where pruning decisions actually get stuck.

Seasonal posts. A page that's dead in September and earns every December fails a naive twelve-month look in July. Judge seasonal pages against their season, and refresh them a month before it starts rather than pruning them at their annual low.

News and announcement posts. "We've moved offices" from 2021 and "Our thoughts on this year's budget" age into clutter by design. These are the easiest retirements on any site, redirect to the about page or the relevant hub, and the lesson going forward is to put dated commentary in formats you expect to expire.

Event and campaign pages. Once the event passes, redirect to the current equivalent or the parent page. Keeping a graveyard of past-event URLs "for the record" serves nobody the record was meant for.

The founder's favorite. Every site has a page someone loves that fails every test. The honest options are the same as for any page, but if it's staying, own the decision as sentiment rather than strategy, and don't let it become a precedent that exempts everything else.

"Can't I just noindex it instead?" A noindex tag asks Google to drop a page from search while leaving it live for humans, and it's the right tool for pages that serve visitors but shouldn't compete in search: thank-you pages, internal-use content, thin utility pages. As a pruning substitute it's usually a half-measure, the page still needs maintenance and still greets whoever finds it, so reserve it for pages with a genuine non-search job, and give true dead weight a real exit.

Aftercare and cadence

Log every removal and merge with its date, then watch two things over the following two months. In Search Console: the redirected URLs' target pages, which should hold or gain, and the site's overall impressions, which should not dip meaningfully, if they do, something on the kill list had a job you missed, and the log tells you what to restore. In your analytics: nothing, ideally, which is the confirmation the pages were as dead as measured.

As routine, pruning earns its place as the small final step of a yearly audit rather than a standalone crusade: fix the fixable, merge the overlaps, and then let the short kill list go without ceremony. Between audits, resist recreational deletion. The compounding that makes content a growing asset lives in old pages quietly gaining authority, and a trigger-happy pruner mistakes "not performing yet" for "never will" at exactly the moment patience was about to pay.

Quick answers on the mechanics

Will deleting pages cause a traffic drop? Not if the test was applied honestly, since you're removing pages that measurably had no traffic to lose. The redirects preserve what little equity the URLs held. The drops people fear come from pruning pages that were actually alive, which is what the impressions check exists to prevent.

How long until I see any effect? Assume weeks to months for the site-quality effect, if it's visible at all at small scale, and treat tidiness, safety, and concentration as the real returns. Anyone promising a traffic spike from deletion alone is reading big-site case studies to a small-site audience.

Should I delete old posts just because they're old? No. Age alone is not a criterion, and old posts with steady traffic are the compounding asset working as intended. Old and dead by every measure is the combination that qualifies.

What about the archive pages, tags, and category URLs my CMS generates? They're pages too, and a blog with 40 posts and 200 thin auto-generated archive pages has a worse quality profile than it thinks. Most small sites do well letting a handful of meaningful categories live and keeping tag sprawl out of the index.

The one-line version

Kill a post only when it fails every test at once: a year old, no impressions, no conversions, no links, no structural job, and a topic you wouldn't choose again, and even then, redirect its URL to the nearest living relative. Fix or merge everything else, prune as the epilogue to a yearly audit rather than a strategy of its own, and spend the drama you save on the pages that deserve it.