Check this page with an assistantOpens a chat asking it to summarise this article and name the evidence behind each claim.
Claude opens with the prompt on your clipboard: Anthropic does not support prefilled prompts on the web, and we would rather copy it than ship a button that drops it.
Decisions first
Compare stable page-query cohorts across equivalent windows, require minimum evidence, separate clicks, impressions, CTR, and position, then classify the likely cause before selecting refresh, snippet test, consolidation, technical fix, or no action.
- Detect at page-query level before rolling up to page.
- Use comparable windows and minimum-impression rules.
- Automate diagnosis packets; keep the intervention accountable.
Define decay as a decision condition
Decay is a sustained loss in a page's ability to perform its intended search job, not any downward movement in clicks.
A page can lose clicks because fewer people searched, Google changed the result layout, the average mix of countries or devices changed, another page inherited the query, a migration changed URLs, or the page itself became less useful. The detector needs the component metrics and context to distinguish those paths.
Start with page-query rows. Aggregate only after each query cluster is classified, because total page clicks can look stable while the commercial cluster disappears and an irrelevant informational cluster grows.
| Pattern | Possible cause | Do not assume |
|---|---|---|
| Impressions down, position stable | Demand or eligibility changed | The copy became worse |
| Position down, impressions down | Competition, quality, technical issue | A rewrite is sufficient |
| Impressions stable, CTR down | Snippet or result-layout change | Rankings collapsed |
| One URL down, another up | Cannibalization or migration | The topic lost demand |
Build a conservative detector
Use two equivalent windows, minimum evidence, and persistence. Flag candidates for review; do not let a noisy week trigger a rewrite.
For stable topics, compare the most recent 28 complete days with the preceding 28 and require a meaningful prior impression count. For seasonal topics, compare year over year or use a longer matched period. Exclude incomplete days and annotate launches, migrations, tracking changes, and known incidents.
A rule can require both an absolute loss and a percentage loss so tiny pages do not dominate. Persist the condition for more than one review. Keep thresholds visible and editable by site; there is no universal percentage that proves decay.
- Normalize page identity across redirects and canonicals.
- Store both raw windows and rule version.
- Segment brand and non-brand where the query data supports it.
- Suppress known migrations and temporary incidents.
- Attach the top gained and lost query clusters.
Turn the alert into a diagnosis packet
The packet should make the first decision cheap: technical incident, demand change, snippet issue, intent drift, competition, cannibalization, or content gap.
Include the metric decomposition, affected query clusters, comparison dates, page changes, indexation and canonical status, internal-link changes, and conversion contribution. Open the current result set for the affected queries to understand intent and result features; Search Console alone cannot show why the result page changed.
Choose the smallest intervention supported by the evidence. A title and description test is cheaper than a full rewrite. Consolidation may be correct when two pages compete. No action is valid when demand normalized after a temporary spike.
Measure the intervention against the diagnosis
A refresh is successful when the intended query job and business outcome recover, not merely when the page's date changes.
Record the hypothesis, exact diff, publish date, expected leading signal, and review window. Technical fixes can be checked quickly; search performance needs enough impressions after recrawl. Avoid comparing a partial post-change window with a full baseline.
Keep a holdout mindset even when a formal experiment is unavailable. Not every flagged page should be changed immediately. A reviewed no-change cohort helps reveal whether the detector is selecting genuine problems or ordinary movement.
Continue the workflow
Backtest the detector on known gains, losses, migrations, and seasonal pages before allowing it to open production work automatically.
- Search Console operator guide: run the weekly evidence review
- content consolidation playbook: choose refresh, merge, redirect, delete, or hold
- indexing diagnostics: separate technical loss from editorial decay
Questions teams ask
- What is content decay in Google Search Console?
It is a sustained decline in a page's intended query performance, visible through the combination of clicks, impressions, CTR, and position after accounting for demand, seasonality, migrations, and URL consolidation.
- What comparison window should a decay detector use?
Twenty-eight complete days versus the preceding twenty-eight is a useful operational start for stable demand. Seasonal topics need matched year-over-year or longer comparable windows.
- Should a decay alert automatically rewrite the page?
No. The same metric pattern can have several causes. Automation should assemble evidence and route the diagnosis; the selected intervention should follow the cause.
Primary sources
The SearchHandled Editorial Team prefers first-party documentation and names the limits of each source. Links were checked on the access date.
- Google Search Console Help: Performance report: clicks, impressions, CTR, and positionAccessed August 29, 2026 · Search Console aggregates and samples some views; it is not a rank tracker.
- Google Search Console Help: Monitor your performance on Google SearchAccessed August 29, 2026
- Google Search Central: Debugging drops in Google Search trafficAccessed August 29, 2026
- Google Analytics Help: Link Search Console and Google AnalyticsAccessed August 29, 2026 · The products use different scopes and will not produce identical totals.
