Refreshing old content: when a rewrite beats a new post

By Matija Konjić

The best-performing content work on most established sites is editing, and almost nobody budgets for it. New posts get calendars, briefs and celebration; the page that quietly slid from position three to eleven gets noticed a year later, if ever. Meanwhile the arithmetic runs the other way: a ranking page that decays holds history, links and indexed trust a new URL would spend months earning, and an afternoon of honest editing routinely recovers what a month of new drafting cannot buy.

This guide is the refresh discipline in full: why pages decay, the triggers that put one on the list, what a real refresh includes, the refresh-or-rewrite decision, and how to run updating as a standing share of production rather than an occasional rescue. It is a core deliverable inside our content writing service, because keeping pages true is writing work of the highest-paying kind.

Why good pages decay

Decay is rarely punishment; it is drift. Facts age: prices change, interfaces get redesigned, laws update, and every stale detail teaches readers the page stopped being tended. Competition arrives: the query that had three serious answers now has ten, and the newest ones learned from yours. And intent moves: the market’s want behind a query shifts, the results page reshapes around it, and a page built for the old want loses to pages built for the new one.

The curve matters because of where it bends. Pages sliding from the top of page one to the bottom, or onto page two, still hold nearly all their accumulated trust, which is exactly why they respond so fast to being made current. The same edit on a page that never ranked does far less, because there was nothing to recover.

The economics follow the curve. A quality refresh costs half to two-thirds of a new piece, lands on a URL that already carries links and history, and pays back inside weeks rather than quarters. On a per-visit basis it is the cheapest organic growth an established site can buy, which is why the fastest-improving programs we see are usually the ones that finally started editing.

None of this argues against new content; it argues for sequencing. New pieces expand the surface, refreshes defend and compound it, and a program running only one of the two is leaving the other half of its returns unclaimed.

The five refresh triggers

A refresh list beats refresh instincts, and five triggers fill the list.

Position slippage is the loudest: any page that fell out of the top five on a term it owned goes on the list automatically. Aged facts are the most common, and the cheapest to spot by skimming your own money pages quarterly. Intent drift shows in the results page more than in your analytics. Out-written is the humbling one, discovered by actually reading whoever took your position. And business drift, the page supporting an offer that changed, is the trigger teams miss longest because nothing in the SEO tooling flags it.

Weight the triggers by stakes as well as signal. A slipping commercial page outranks a slipping informational one on any sane list, and a stale claim on a money page is an urgent defect rather than a backlog item, because the reader it misleads was about to become a customer. The list is a queue, and the queue has a front.

Finding candidates in ten minutes

Search Console does the finding: filter to pages whose clicks fell year over year, sort the queries sitting at positions eight through twenty, and the refresh list writes itself. The same review inside a content audit produces the annual version; the quarterly ten-minute pass keeps the list warm between audits.

Set a floor for what earns a slot, because refreshing everything is how nothing gets refreshed. Pages with real history, links or commercial roles qualify; pages that never ranked and never mattered route to the audit’s merge and kill verdicts instead. The refresh list is for recoverable value, and guarding that definition keeps the queue short enough to actually drain.

What a real refresh includes

A refresh is re-reporting, never repainting. The working checklist:

• Re-read the results page first. The current winners show what the intent demands now, which sections the market added, and what your page is missing.
• Update every fact. Prices, versions, screenshots, examples, dates, links, checked against sources rather than memory.
• Rebuild the weak sections. Thin answers get depth, missing subtopics get added, and anything the winners cover that you skip gets an honest treatment.
• Cut what aged badly. Dead tactics, retired tools and expired context leave, because keeping them taxes everything around them.
• Rewire the links. New internal links in from recent pieces, links out to the pages that now exist, anchors still natural.
• Re-run the editorial bar. The refreshed piece passes the same claims-and-sources exam a new piece would.

Done that way, the page is current in substance rather than costume, and the date change it earns is true. Changing the date without the work is the industry’s least convincing trick, and it burns exactly the trust a refresh exists to rebuild.

A worked refresh, in one paragraph

A pricing guide published two years ago sits at position nine, down from three. The re-read shows winners now leading with a comparison table and covering two pricing models that did not exist at publication. The refresh updates every number against current sources, adds the two missing models, replaces the dead screenshots, cuts a section about a retired tool, links in from the three newest related posts, and re-states the summary the answer engines keep lifting from rivals. Total cost: an afternoon. The page re-enters the top five inside six weeks, on authority it had held all along.

Multiply that example across a mature archive and the quarter’s plan writes itself: ten such afternoons often outperform ten new posts on every metric that pays, and they finish sooner, which is the rarest sentence in content planning.

Refresh, rewrite or retire

Not every aging page deserves the same intervention, and the decision has three exits.

Refresh on the same URL when the bones are right: structure matches intent, the angle still holds, and staleness is the main crime. Rewrite on the same URL when the topic still earns a page but the execution is wrong for today’s results, keeping the address so history transfers to the better piece. Merge or retire when the topic itself stopped deserving a page, with redirects to the closest survivor. The URL is the asset; the rule across all three exits is that its history never gets abandoned casually.

The instinct to publish fresh instead of fixing old is worth naming, because it decides careers as well as pages: new URLs feel like progress and edits feel like chores, so calendars fill with the expensive option. Treat the decision as arithmetic instead. History plus half-cost against zero plus full-cost is rarely a close call, and the chart from the last worked example is what the arithmetic looks like when somebody finally runs it.

Running refreshes as a rhythm

The programs that compound treat updating as a standing share of production: one refresh slot for every four or five new pieces, filled from the trigger list, briefed like any other assignment. The share rises with the archive’s age, and mature sites often profit from flipping the ratio entirely for a quarter, because their fastest available wins are all in the back catalog.

Ownership makes the rhythm stick. Every significant page gets a revisit date at publication, pillar pages yearly at minimum, fact-heavy pages sooner, and the calendar surfaces them without anyone having to remember. That single habit converts refreshing from a rescue activity into maintenance, which is what it always should have been.

Fold refresh entries into the same roadmap and briefs as new work, with the same editorial pass at the end, so quality never depends on which kind of assignment a writer drew. A refresh brief is shorter, what changed, what the winners now cover, what must be verified, and it earns the same twenty minutes a new brief gets, for the same reason: undirected editing wanders exactly like undirected drafting, and a wandered refresh can damage a page that was merely stale.

Measuring recovery

Refresh results read fast by content standards. Recrawl comes within days, movement on the recovered terms typically shows inside four to eight weeks, and the comparison that proves the work is simple: clicks and positions for the page, ninety days before against ninety days after, annotated with the ship date. Report refreshes as their own line beside new content, because their economics are different and better, and stakeholders should see that.

Watch the secondary effects too, because refreshed pages lift their neighbors. Recovered pages send internal authority back through the cluster, their fresh sections win new long-tail queries the old version never touched, and answer engines pick the updated passages up on the next crawl. The before-and-after on the page understates the before-and-after on the group, which is one more reason to measure clusters rather than posts.

Expect a hit rate rather than perfection. Most well-chosen refreshes recover meaningfully; some confirm the topic’s demand moved on, which routes them to the merge-or-retire exit with evidence in hand. Both outcomes beat the alternative every mature archive otherwise defaults to, which is decaying politely in public while the calendar ships something new beside it.

The discipline also compounds reputationally. A site whose pages are reliably current gets cited, bookmarked and returned to, and readers who got burned once by stale advice remember the domain that burned them for a long time. Freshness is a trust signal to humans before it is anything to machines, and the refresh rhythm is simply what keeping your word looks like at archive scale. That is the whole case: cheaper than new, faster than new, and it protects everything the site already earned. If the back catalog is where your fastest wins are parked, that is exactly the work we would start with.