The content refresh system takes a page that used to perform and works out why it stopped, then rebuilds it. The diagnosis matters more than the rewrite, because a page losing clicks while holding its position has a different problem from one that’s been overtaken.
Built in-house. It is my own tooling rather than a reseller’s dashboard, which is why it checks the things I act on rather than everything a generic tool can measure.
What it does
It pulls the page’s own history from Search Console, compares it against what currently ranks, identifies what’s changed on the results page itself, and produces a rebuilt version with the parts worth keeping retained.
Roughly 60 to 70 per cent of the page ends up genuinely new, with 30 to 40 per cent kept. That ratio isn’t a rule, it’s what a real refresh usually comes to. A rewrite that keeps almost nothing is a new page, and one that keeps almost everything is a tidy-up.
The diagnosis comes first
| Pattern | What it usually means |
|---|---|
| Position held, clicks fell | The results page changed around you. An AI Overview, more ads, or a feature pushing you down |
| Position fell steadily | Competitors improved, or the page has gone stale against a topic that moved |
| Position fell sharply on one date | An algorithm update, a site change, or something technical |
| Impressions fell too | Demand itself has fallen, which no rewrite fixes |
| Clicks fell, CTR collapsed | Often the title being rewritten in results, so check the live listing |
That last one is worth its own note. Before calling a low click-through rate a content problem, check the title actually showing in the live result. Search engines rewrite titles, and a template leaking something odd into the listing looks exactly like a content problem from inside Search Console.
The fourth row is the one clients least expect. If the searches have declined, the page is holding its share of a smaller market, and rewriting it changes nothing. Saying so is more useful than producing work.
What the rebuild covers
- What the ranking pages now answer that this one doesn’t
- Whether the answer sits near the top in a form that can be quoted
- Factual claims rechecked against primary sources, with the access date recorded
- Anything now wrong, including retired features, moved documentation and old figures
- Internal links, both in and out, since a declining page is often one that lost its links
- Structure, headings and schema
Example output
PAGE: /guides/does-lube-expire/ HISTORY: Position stable at 3.1. Clicks down 41 per cent over 6 months. DIAGNOSIS: An AI Overview now appears on the main search. Position held, the click didn’t. ACTION: Rewrite the opening as a direct 50 word answer, add the detail the overview omits, add FAQ schema. KEEP: The storage table and the shelf life figures, which are the strongest part.
What I review
Every diagnosis is checked against the live search result, not just the data. Search Console tells you what happened to your page and nothing about what happened to the page it appears on.
Sources are rechecked rather than carried over, because moved or retired documentation is one of the most common ways an old guide becomes quietly wrong.
Limitations
- A refresh can’t recover lost demand. If fewer people search, fewer people arrive.
- Results take time and aren’t guaranteed. Recrawling and reassessment happen on the search engine’s schedule.
- Some pages shouldn’t be refreshed. Merging into a stronger page or removing it is often the better answer.
Services it powers
The content refresh system powers content refreshes and feeds content quality and E-E-A-T work. See the rest of the tools on the tech stack page.
FAQ
Why has my page lost traffic without losing rankings?
Usually because the results page changed around it. An AI Overview, more ads or a new feature can push a stable position further down the screen, so the ranking holds and the clicks don’t. Checking the live result is the only way to confirm it.
How much of a page changes in a refresh?
Usually somewhere between 60 and 70 per cent, with the strongest 30 to 40 per cent kept. That isn’t a target. A rewrite keeping almost nothing is really a new page, and one keeping almost everything is a tidy-up rather than a refresh.
Should every declining page be refreshed?
No. Some should be merged into a stronger page and some removed. And where the decline is in search demand rather than in performance, the page is holding its share of a smaller market, so rewriting it changes nothing worth having.
See the whole toolkit
The tech stack page lists every tool, and the free assessment identifies pages worth refreshing.