Why We Kill Articles That Don't Earn Their Place

Most content sites operate on a one-way ratchet: publish, then never look back. An article goes live, it ages, it stops getting traffic, and it sits there forever as a tiny liability nobody wants to touch. We run the opposite policy. Every article we publish is on probation, and a fair number of them get cut.

That sounds harsh for a site that depends on having things to read. It isn’t. Killing weak articles is one of the few editorial decisions that helps almost everything else on the page. Here is how we think about it, what triggers a review, and what actually happens when an article fails to earn its keep.

The cost of keeping a dead article alive

A published article is not free to keep around. It carries ongoing cost whether or not anyone reads it.

The first cost is crawl and ranking dilution. Search engines have a finite appetite for any one site. When a third of your URLs are thin, outdated, or duplicative, they compete with your good pages for the same attention and the same internal link equity. You end up with ten mediocre pages splitting the authority that two strong pages could have concentrated.

The second cost is link rot. We’re an affiliate and tools site, which means a meaningful share of our outbound links point at products that change pricing, rebrand, or shut down. An article reviewing a tool that pivoted eighteen months ago isn’t neutral — it’s actively wrong, and a reader who follows a broken recommendation doesn’t blame the tool. They blame us.

The third cost is trust math. Every article is a sample of your judgment. A reader who lands on one stale, padded, clearly-AI-extruded post will discount the next three good ones they read. You don’t get credit for the average; you get judged by the worst thing they happen to find.

So the question we ask isn’t “is this article bad?” Most aren’t outright bad. The question is whether each one still earns the cost of keeping it indexed, linked, and presented as something we stand behind.

How we decide what to kill

We review articles on a rolling basis, and each one lands in one of three buckets: keep, revise, or kill. The decision isn’t a vibe. It runs against a short list of measurable checks.

The two columns that decide most cases are uniqueness and effort-to-fix. If an article overlaps a stronger post we’ve already published, the right move is almost never to keep both — it’s to merge the useful parts into the stronger one and redirect the weaker URL to it. That consolidates ranking signals instead of splitting them, and the reader still lands somewhere useful.

The effort-to-fix column is where honesty matters. If fixing an article means a full rewrite, you are not revising — you are writing a new article and pretending it’s maintenance. Sometimes that’s worth it for a high-intent topic. Often it isn’t, and the disciplined call is to delete the old one and spend that writing time on a topic that’s actually live.

We track all of this in a single board so the decision has a paper trail. Each article has its traffic trend, last accuracy check, link status, and a one-line disposition. When something gets killed, the reason is logged next to it — which keeps us honest and stops us re-publishing the same dead-end topic six months later.

What killing actually looks like

Deleting an article well is not the same as hitting delete. A URL that 404s after it built any links or rankings is wasted equity and a worse experience than the stale page was.

When we kill an article, one of three things happens to its URL. If there’s a stronger article on a closely related topic, we 301-redirect the dead URL to it, passing whatever ranking signals it accumulated to a page that can use them. If the topic is dead but the URL had real inbound links, we may keep a short, honest stub explaining the change. Only when a URL has no value at all — no traffic, no links, no relevance — do we let it go to a clean 404 or 410.

We also resist the temptation to kill on a single bad week. Traffic is noisy, seasonality is real, and a redesign or an algorithm update can briefly tank a page that recovers fine. The kill threshold is two flat quarters, not one slow month. The cost of keeping a borderline page another quarter is small; the cost of redirecting away a page that was about to recover is not.

The payoff is concentration. Pruning isn’t about having less — it’s about every remaining article being something a reader benefits from finding and we’re comfortable having our name on. A smaller catalog of pages that each earn their place beats a sprawling one where you quietly hope nobody clicks the wrong link.