Every content library decays. Rankings slip, information ages, better competitors publish. The teams that compound treat refreshing as a programme with rules — not a panic response to a traffic dip. Here is the framework.
Separate a search change from a page change
Before editing an older page, record what actually moved. Search Console can show that a URL lost impressions, lost clicks or moved down for a particular set of queries. Those are different problems. A drop in impressions may reflect lower demand, while stable impressions and fewer clicks may suggest a changed search result or weaker title. A drop concentrated on one query may require a smaller edit than a loss across the whole topic.
Compare matched periods where possible. A seasonal page can look broken when October is compared with a summer peak. Check whether a redesign, redirect, title edit or internal-link change happened near the decline. Then inspect the current results: did Google start showing tools where it once showed guides, or did a competitor publish a better answer? This diagnosis should be in the brief before a writer begins.
Our own site offers a structural example. A Leeds head query was appearing on several pages, including About and Contact, although the homepage was intended to answer it. Refreshing each page with more city phrases would have deepened the overlap. We chose one owner and changed the supporting pages to link to it in context. That is a content decision as much as a technical one.
Write a refresh brief that can be reviewed
The brief should name the current URL, primary intent, important supporting queries, existing sections to keep, facts that need verification, new evidence available, and links that should change. It should also name the person who can approve claims. For regulated services, that may be a solicitor, clinician or financial adviser; an SEO writer should not guess at compliance wording.
When a page already earns impressions for several relevant queries, preserve the useful answer near the top. Add material that resolves an unmet decision: a clearer comparison, an example with dates, a process explanation or a limitation. Remove stale claims and unsupported superlatives. A new modified date should reflect a real editorial change, while the original publication date should remain truthful.
Set a review date before publishing the refresh. Recheck the same query and page group after enough data accumulates, and annotate any other changes made at the same time. If the page does not improve, examine intent and authority before rewriting again. A site with weak external links cannot reliably edit its way into a competitive result, even with a strong page.
Keep the before and after versions of the page in your change log. That record lets a later editor see which facts were corrected, which sections moved and why the page was kept instead of merged. It also makes future diagnosis faster when several people work on the same content library.
Catching decay at position 5 is a refresh; catching it at position 25 is a rebuild.
Detect decay before it becomes collapse
Set a quarterly review comparing every significant page’s clicks and average position against the prior period. Decay has a signature: position drift from 3 to 5 to 8 over months, often while the page looks "fine". Catching it at position 5 is a refresh; catching it at position 25 is a rebuild.
Prioritise by value at risk: pages with conversion history or link equity get attention first. A decaying post that never converted anyone may deserve consolidation instead of rescue.
Diagnose why before touching anything
Four distinct diseases produce the same falling line. Outdated information: facts, screenshots, prices, or years that scream neglect. Intent drift: the SERP changed shape — what wanted a guide now wants a tool or a comparison. Competitive displacement: someone simply published better coverage. Cannibalization: your own newer post is splitting the query.
Each has a different cure, and applying the wrong one wastes the effort. Read the current SERP for the target query before opening the editor: what Google rewards today is the spec you’re writing to.
Refresh, consolidate, or redirect — the decision rules
Refresh when the page still matches intent and holds equity: update facts and examples, close the coverage gaps competitors opened, strengthen E-E-A-T signals (author, evidence, dates), and improve internal links. Keep the URL; update the modified date honestly.
Consolidate when multiple weak pages split one intent: merge the best material into the strongest URL and 301 the rest. Redirect or prune when the topic is dead or was never yours to win — thin pages with no traffic, no links, and no strategic role dilute sitewide quality signals.
Execute refreshes like releases
Batch refreshes into a monthly cycle with briefs, owners, and deadlines — the same discipline as net-new content. A refresh brief lists what changed in the SERP, which sections to add or rewrite, which internal links to add, and what success looks like at 30 and 90 days.
Resist the temptation to change the things that were working: if the page ranks for forty long-tail variants, wholesale rewrites can vaporize them. Diff carefully; expand more than you replace.
Measure recovery and feed the learning back
Track position and clicks at 30/60/90 days post-refresh. Successful patterns — "adding comparison tables recovered every commercial-intent post" — become standing rules for the next batch. Failures redirect effort: a page that doesn’t respond to a competent refresh is usually fighting an intent shift or authority gap that editing can’t fix.
Mature programmes typically find refresh work outperforms net-new publishing on ROI: the page already has age, links, and indexed history. It just needs to deserve its ranking again.
The useful takeaway
Review quarterly, diagnose the specific disease before treating, choose refresh/consolidate/prune by rule, execute in disciplined batches, and measure recovery. Old content with equity is the cheapest growth most sites are ignoring.
