Content Freshness: When to Update Old Posts
Somewhere in your archive is a post from three years ago that still gets a trickle of search traffic, ranks on page two for a term you'd love to own, and contains at least one screenshot of an interface that no longer exists. That post is worth more than the next three posts you're planning to write — if you treat it right.
Updating existing content is usually the highest-ROI move available to a small site, for a simple reason: an old post has assets a new post doesn't. It's already indexed, already has internal links, possibly a backlink or two, and an established history with search engines. A refresh starts from position 14; a new post starts from nowhere. But "update old posts" has also become cargo-cult advice — people changing dates and shuffling paragraphs and wondering why nothing moves. So let's separate the mechanism from the ritual.
What "freshness" actually means to a search engine
Google's handling of freshness is query-dependent, and understanding that saves you from updating the wrong things. Some queries deserve fresh results — "best accounting software," "iphone camera comparison," anything with a year in it, news, prices, regulations. For these, recency is a genuine ranking input, and a stale post decays no matter how good it was. Other queries are freshness-indifferent — "how to fold a fitted sheet," "what is compound interest." The old answer is still the answer, and slapping a new date on it earns nothing.
Two implications:
- Freshness is about the content, not the timestamp. Search engines compare versions of your page. A date change with cosmetic edits is detectable and ignorable. Substantive changes — new sections, corrected facts, updated recommendations — are what register.
- Your archive splits into decay-prone and evergreen posts, and they need different maintenance schedules. Product roundups and anything tied to versions, prices, or laws decays in months. Conceptual explainers can coast for years.
Triage: which posts to update first
Don't refresh chronologically or by gut feeling. Pull two data sources — Google Search Console (Performance, last 12 months, compare to previous 12) and your analytics — and sort your posts into four buckets:
| Bucket | Signature | Action |
|---|---|---|
| 1. Decliners | Had meaningful traffic, clearly trending down | Refresh first — proven demand, fixable decay |
| 2. Almost-rankers | Positions 5–15 for a valuable query, decent impressions, weak clicks | Refresh second — smallest push, biggest payoff |
| 3. Steady earners | Stable traffic, still accurate | Leave alone; check facts annually |
| 4. Dead weight | No traffic, no rankings, no links, redundant or thin | Merge into a stronger post (with a 301) or prune |
The almost-rankers are the secret weapon. Moving from position 12 to position 5 can multiply a page's clicks several times over; moving from 45 to 30 changes nothing anyone will notice. Search Console makes these easy to find: filter for queries where your average position is between 5 and 15 and impressions are respectable but CTR is low.
A reasonable cadence for a small site: pick a few posts a month from buckets 1 and 2. That's one afternoon, and over a year it means your best posts stay sharp — which for most small sites covers the pages earning most of the organic traffic anyway.
What a real refresh looks like
A refresh worth doing changes the substance. Work through this sequence:
1. Re-check the query before touching the post
Search your target term and look at what ranks now. Sometimes intent has drifted: a query that used to return how-to guides now returns comparison lists, or vice versa. If the winning format changed, your update needs to change format, not just facts. This five-minute check prevents the most common refresh failure — polishing a post into a better version of the wrong thing.
2. Fix everything that's wrong
- Outdated facts, prices, version numbers, screenshots, and dated "as of" phrasing.
- Dead outbound links (a crawler or the free Broken Link Checker finds them) and internal links pointing at redirects.
- Recommendations you no longer stand behind. This is the credibility one — a post recommending a plugin abandoned two years ago tells every knowledgeable reader you don't maintain your content.
3. Close the gaps
Read the People Also Ask questions and the top three ranking pages (for structure, never for copying). What questions do they answer that you don't? Add the missing sections. Then check Search Console's query list for the page itself — if it's getting impressions for a question you only half-address, that's a section begging to be written. This is often where an update adds 300–600 genuinely new words.
4. Cut as ruthlessly as you add
Refreshing is not inflating. Delete the meandering intro from an older writing style, the tangent that never earned its place, the third example making the same point. A post that goes from 2,400 mediocre words to 1,800 sharp ones has been improved, whatever the word-count worshippers say.
5. Rewire it into the site
Old posts predate your newer content, so they can't link to it — until you make them. Add internal links from the refreshed post to relevant newer pieces and, crucially, from newer high-traffic pieces back to it. If you've organized content into topic clusters, this is the moment to wire the post into its cluster properly.
6. Update the date honestly
If you made substantive changes, update the modified date and let your theme display "Last updated." Keep the original publish date in the record — longevity plus maintenance is a trust signal, and "Published long ago, updated recently" reads better to humans than a suspiciously young post with 40 comments from three years ago. Do not change dates on posts you haven't substantively changed. Readers notice, and it burns trust you can't easily rebuild.
7. Keep the URL
The URL holds the page's accumulated equity. Don't change it to add a keyword or remove a year — the 301 works, but you're trading a settled asset for a small disruption with no upside. (Exception: if the URL contains a year like /best-tools-2023/, learn the lesson once, redirect to a year-free URL, and never put years in slugs again.)
When to merge or delete instead
Some posts shouldn't be saved. Signs a post belongs in the merge-or-prune pile:
- It competes with your own better post. Two posts targeting the same query split your signals. Merge the weaker into the stronger, keep the best material from both, 301 the loser.
- It was never good and has no traction. A thin old post with zero traffic and zero links isn't an asset to polish; it's clutter. Improve it only if the topic genuinely fits your strategy — otherwise prune.
- The topic no longer relates to your business. Authority is topical. Thirty off-topic posts from a previous content strategy dilute what your site is about.
Pruning feels destructive and almost never is. Removing pages that earned no traffic can't lose you traffic; it concentrates crawl attention and topical focus on the pages that matter.
What to expect afterward
Refreshed posts typically show movement within two to eight weeks — faster than new content, because the page isn't starting from zero. Track each refreshed URL in Search Console (compare the 28 days after vs. before, position and clicks). Expect maybe half your refreshes to show a clear win, some to do nothing, and the occasional one to dip — usually the ones where the real problem was competition, not staleness. That hit rate is fine. The compounding effect of a maintained archive versus a decaying one is the actual prize, and it shows up in the annual totals rather than any single post's chart.
The takeaway
Your archive is either an appreciating asset or a decaying one; the difference is a few hours a month applied to the right posts. Triage by data, refresh substance instead of timestamps, prune without sentimentality, and keep your steady earners honest once a year. New posts expand your territory — updates defend it. Most small sites are much better at the first than the second, which is exactly why the second pays so well.
Photo: Nothing Ahead on Pexels