decay-scanner

name	decay-scanner
description	Detect content decay — pages that have lost organic traffic versus a past peak. Use when the user says "find decaying pages", "what traffic did I lose", "content decay audit", "why is traffic dropping", "which posts are declining", or pastes two GSC exports (current vs. an earlier/peak period) and asks what slipped. Works standalone from two GSC CSVs.

Decay Scanner

Compare a current GSC export against an earlier peak-period export and flag pages that have lost significant organic clicks. Decay is the quiet killer of content sites — pages that once ranked slowly bleed traffic to fresher competitors or algorithm shifts. Catching it early makes refreshes cheap.

Input

Two GSC exports, ideally at the Page level (or Query + Page):

• Current period (e.g. last 28 days, or last month).
• Peak / baseline period (the page's historical best, or same period last year to control for seasonality).

Expected columns in each: Page (or Query + Page), Clicks, Impressions, CTR, Position. Ask the user to confirm the date ranges and that the window lengths match (comparing 28 days to 90 days is meaningless — normalize or re-export). If only one file is provided, ask for the second; you cannot scan decay from a single snapshot.

Steps

1. Align periods. Confirm both windows are the same length and note them in the output. If lengths differ, normalize to clicks/day before comparing and say so.
2. Join on Page. Match URLs across both exports (normalize trailing slashes / fragments).
3. Compute deltas per page: Δclicks, % change, Δimpressions, Δposition, ΔCTR.
4. Flag decay where clicks dropped ≥20% from peak AND the page had meaningful baseline traffic (e.g. ≥50 peak clicks — skip noise). State thresholds.
5. Assign severity:
   • Critical — ≥60% drop, or lost a top-3 position, or high absolute click loss.
   • High — 40–60% drop.
   • Moderate — 20–40% drop.
6. Diagnose likely cause from the available signals (label your confidence):
   • Position held, impressions/CTR fell → likely SERP change (new SERP feature, AI Overview, competitor took the snippet) or seasonality.
   • Position dropped, impressions fell → likely competition / algorithm (someone outranked you, or a core update reweighted the page).
   • Impressions steady, CTR dropped → likely SERP feature eating clicks or a title/meta that aged out.
   • Same period last year shows the same dip → seasonality, not a problem to chase.
7. Recommend action per page tuned to cause: refresh + republish (stale content), title/meta rewrite (CTR loss), rebuild for the new dominant intent (SERP shifted), reclaim the snippet, add internal links, or "monitor — seasonal, no action."

Output

A decay report table sorted by severity then absolute clicks lost:

Page	Peak clicks	Current clicks	% change	Δ Position	Likely cause	Severity	Recommended action

Then a short narrative: total clicks lost across all decaying pages, the 3 highest-value refreshes to prioritize (biggest absolute loss × easiest fix), and explicitly separate seasonal dips (no action) from real decay (act now) so the user doesn't waste effort. If a query-level export is available, name the specific queries each page lost so the refresh is targeted.