Key takeaways
- A blog post's rankings typically start decaying 6–12 months after publishing, visible in Search Console as falling impressions and average position.
- Three signals decide the call: whether a related URL already ranks, how stale its facts are, and whether the topic exists in the library at all.
- Updating wins when a post already holds a page-one or page-two position, because it inherits backlinks and crawl trust a new URL has to earn from zero.
- Writing new wins when the topic doesn't exist yet, the offer changed enough to need a rebuild, or a competitor's format can't be patched into the old post.
- AI can scan dozens of posts for decay signals in minutes, but deciding which flagged posts matter this quarter still depends on judgment only the business has.
Publishing more posts doesn't fix a content engine that's leaking traffic from the archive you already built. The real question is where the next hour of writing time — yours, your team's, or an AI assistant's — produces the most qualified leads: fixing a page Google already trusts, or building a new one from zero authority.
That's a measurement problem, not a preference. The answer changes post by post, and getting it right is what separates a content program that compounds from one that just accumulates word count.
Old blog posts decay on a predictable timeline
A blog post's rankings and traffic follow a decay curve that typically starts 6–12 months after publishing, as competitors publish newer material and search intent shifts under the topic. This isn't a guess — it shows up directly in Google Search Console as a slow decline in impressions and average position for a URL that used to hold steady.
For an owner running one content workstream with no dedicated analyst, that decay is invisible until it's severe. A post that drove 40 organic visits a month in year one can fall to 8 by year three, not because it got worse, but because everything around it got newer. Auditing the archive on a schedule — quarterly, not "when someone remembers" — is what catches the decline while a light update can still reverse it, instead of after the post has fallen off page two entirely.
The refresh-or-write decision comes down to three signals
Three signals decide whether a topic needs an update or a new post: existing rankings and traffic on a related URL, how stale the facts or examples in that URL are, and whether the topic is even covered in your current library at all. Run those three checks before writing anything, because skipping them is how budgets get spent rebuilding pages that didn't need it.
If a URL already exists and ranks anywhere on page one or two, check it first. If the facts, pricing context, screenshots, or examples are outdated but the structure and search intent match are still sound, it's an update. If no URL exists for the topic, or the existing one targets a completely different search intent, it's a new post. Running this triage against ten or twenty candidate topics a quarter — instead of defaulting to "write something new" — is the single highest-leverage habit in a lean content operation.
Updating outperforms writing new when a post already ranks
Updating wins when a post already holds a page-one or page-two position and only needs current data, a sharper answer near the top, or expanded depth to match what's now ranking above it. The page has already earned backlinks, internal links, and crawl trust — assets a brand-new URL has to build from scratch, which is typically the slower and more expensive path.
Picture a two-person sales team running a $3k/month ad budget with one part-time marketer handling content on the side. A 2022 post on "how much does a kitchen remodel cost" still ranks position 9, but it cites 2022 material prices and doesn't mention financing options the business now offers. Rewriting the pricing section, adding a financing paragraph, and refreshing the publish date is a two-hour task that can move that post from position 9 to position 4. Writing a brand-new post on the same query would compete against the URL that's already there — including, ironically, your own.
How do you know if an old post is worth updating?
A post is worth updating if it still gets any organic traffic, ranks anywhere in positions 4–30 for a query tied to a service you sell, or covers a topic prospects still ask about on sales calls. Those three checks take fifteen minutes per post in Search Console and your CRM notes, and they separate posts worth investing in from ones better left alone or retired.
Posts that fail all three — zero traffic, no ranking, and a topic nobody asks about anymore — aren't refresh candidates. Either merge them into a stronger page, redirect them, or leave them alone rather than spending rewrite time on a page with no underlying demand. The goal isn't to touch every old post; it's to find the ones where a small edit unlocks demand that's already searching.
AI speeds up the content audit and rewrite, not the judgment call
AI can scan a hundred old posts in minutes and flag decay signals — thin sections, outdated dates, missing entities, competitor pages that now outrank yours — far faster than a person clicking through Search Console one URL at a time. What it can't do is decide which of those flagged posts actually matters to your business this quarter; that call still depends on which services you're pushing, which cities you're targeting, and what your sales team is hearing on calls.
The failure mode that makes AI content sound generic is skipping straight from "flag this post" to "regenerate this post" without adding anything specific. A workflow that avoids it: feed the AI assistant the existing post, the current top-five ranking pages for that query, and two or three real examples from your own jobs or client conversations, then ask it to identify what the ranking pages cover that yours doesn't — not to rewrite the whole piece. You edit in your own numbers, your own constraints, and the specific trade-offs your customers actually face. The AI does the comparison work; you supply the parts that make it sound like your business instead of every other business that ran the same prompt.
Refresh work compounds
When does writing a new post beat updating an old one?
Writing new beats updating in three cases: the topic doesn't exist anywhere in your library yet, the service or offer has changed enough that the old page would need a full rebuild rather than an edit, or a competitor has taken over the ranking with a fundamentally different angle the old post can't be patched into. Trying to force an update onto any of these wastes the time you saved by not starting from a blank page.
If you just added AI phone answering to your service list and have never written about it, that's a new post — there's no existing asset to build on. If your old "emergency plumbing repair" post was written around dispatch times you no longer offer because you've automated intake, patching in new sentences won't fix a structure built around a workflow that no longer exists; write it fresh. And if a competitor's post now answers the query with a comparison table or calculator your old post can't accommodate without a rebuild, match the format with a new page instead of bolting a table onto old prose.
Does refreshing an old post cost less than writing a new one?
Refreshing an existing post is typically less expensive than publishing a new one, because it inherits backlinks, internal links, and search-engine trust that a brand-new URL has to earn from a standing start. That head start typically means faster ranking movement for less writing and editing time, which matters directly when your content budget is one part-time marketer's hours plus whatever AI-assisted drafting you're running.
The trade-off is that refreshes have a ceiling — you can't refresh your way into covering a topic you've never written about, and a library with real gaps in it needs new posts regardless of how efficient the refresh workflow gets. The practical split for most lean operations: roughly 60–70% of a quarter's content hours on refreshing what already ranks with fixable gaps, and the rest on new posts for topics the archive doesn't cover at all. Track posts by traffic and query gap in the same sheet, and the split becomes obvious instead of a guess.
Build the refresh into a repeatable content system
A repeatable audit-and-refresh cycle outperforms one-off "should I update this" decisions made post by post, because decay doesn't stop after one edit — it starts again the moment a competitor republishes. Put a quarterly cadence in place: export Search Console data, run it through the three-signal triage above, and route each post to update, retire, or leave alone before touching a single sentence.
That cadence is also where AI earns its keep at scale — running the same audit prompt against fifty URLs instead of five, so the constraint on how much of your archive gets reviewed each quarter is time, not tooling. See how this fits into a full editorial, social, and SEO system on the content services page, or look at past project work for how the audit-to-refresh cycle has played out for other service businesses. If you're not sure which of your own posts are worth saving, that's a fifteen-minute conversation — book time and bring your Search Console export.
Prefer it done for you? This playbook is our Content engine: see how we run it for clients →
Frequently asked questions.
How do I know if an old blog post is worth updating?
Check three things in about fifteen minutes: whether the post still gets organic traffic, whether it ranks anywhere in positions 4–30 for a relevant query, and whether prospects still ask about that topic on sales calls. If a post fails all three, it isn't a refresh candidate — merge it, redirect it, or leave it alone.
Is updating an old post faster than writing a new one?
Yes, in most cases. Refreshing an existing post is typically less expensive than publishing a new one because it inherits backlinks, internal links, and search-engine trust a brand-new URL has to earn from a standing start, which usually means faster ranking movement for less writing time.
When should I write a new post instead of updating an old one?
Write new when the topic doesn't exist anywhere in your library yet, when the service or offer changed enough that the old page needs a full rebuild rather than an edit, or when a competitor now ranks with a fundamentally different format the old post can't accommodate.
How often should I audit old blog posts for updates?
Run the audit quarterly, not on an ad hoc basis. A quarterly cadence catches decay while a light update can still reverse it, and it's frequent enough to route each post to update, retire, or leave alone before the archive falls behind.