SEO-08 · SEC. 07 SEO & GEO

Content Decay Audit: Which Pages Are Losing Rank and Why

Diagnose why specific pages are sliding in traffic before you spend time rewriting them.

FORMAT: workflow
DIFFICULTY: intermediate
TIME: 15 min
TOOLS: universal
MODELS: any
COPIES: 0 so far

When to use this

Monthly or quarterly, when Search Console shows a set of pages trending down in clicks or average position. Use this to sort "needs a refresh" from "lost to a real competitor" from "cannibalized by our own page" before spending writing time on the wrong fix.

The pattern

Pastes as plain text

Below the line is Search Console / analytics data for pages on my site,
showing URL, clicks, impressions, and average position for the current
period vs. the prior period of the same length. Treat the window as the
last 90 days vs. the prior 90 unless the data says otherwise.

For each page showing a decline (clicks or position down at least 15%):
1. Classify the likely decay cause using only the data given:
   - STALE: impressions flat or down, position down — content likely
     outdated versus fresher competing results
   - CANNIBALIZED: impressions flat, but clicks/position down while
     another of my own pages on a similar topic gained — check the list
     for a page covering similar ground
   - SERP CHANGE: impressions themselves dropped sharply — likely a new
     SERP feature (AI Overview, featured snippet) taking the click, not
     a ranking loss
   - SEASONAL: this pattern matches a prior-year dip for the same period,
     if that's inferable from the data
   - UNKNOWN: doesn't clearly match the above — say so rather than forcing
     a category
2. For STALE pages, name the specific signal of staleness if inferable (a
   date in the title/content, a "last updated" mismatch, outdated stats).
3. For CANNIBALIZED pages, name the specific other URL competing for it.
4. Output a prioritized refresh list: URL | Decay type | Evidence | Clicks
   lost (absolute number) | Recommended action, sorted by clicks lost.

Paste the traffic data below this line:

Real example output

Page: /blog/best-invoicing-software-2024
Current: 340 clicks, 12,000 impressions, avg position 8.2
Prior: 890 clicks, 11,800 impressions, avg position 4.1
-> STALE. Impressions held steady but position dropped 4 spots. Title and
H1 both say "2024" and it's now mid-2026 — clear staleness signal, likely
losing to freshly-dated competing roundups.
Clicks lost: 550.
Action: Update title/H1 to current year, refresh the comparison table with
current pricing, republish with a real updated date.

Page: /features/invoice-reminders
Current: 210 clicks, 6,400 impressions, avg position 6.5
Prior: 480 clicks, 6,300 impressions, avg position 3.9
Also in dataset: /features/automated-payment-reminders launched 4 months
ago, current 410 clicks, avg position 2.1, same core topic.
-> CANNIBALIZED by /features/automated-payment-reminders. The newer page
is winning the same query.
Clicks lost: 270.
Action: Consolidate — canonicalize /features/invoice-reminders to the
newer page, or differentiate it to target a distinct sub-intent instead
of competing head-on.

Page: /docs/getting-started
Current: 1,100 clicks, 41,000 impressions, avg position 2.8
Prior: 1,300 clicks, 61,000 impressions, avg position 2.6
-> SERP CHANGE. Position barely moved but impressions dropped 33% —
consistent with an AI Overview now answering "how do I get started with
the product" directly in the SERP without a click-through.
Clicks lost: 200.
Action: Not a content problem. Strengthen the page's citability so it's
the source the AI Overview pulls from (see citation-worthiness audit).

Prioritized refresh list:
URL                                    | Type         | Clicks lost
/blog/best-invoicing-software-2024     | STALE        | 550
/features/invoice-reminders            | CANNIBALIZED | 270
/docs/getting-started                  | SERP CHANGE  | 200

Why it works

"Traffic is down" has at least four different root causes and each one needs a different fix — rewriting a page that's actually being cannibalized by your own newer page wastes the effort twice. Classifying decay type from the data before touching copy stops you from refreshing a page that was never stale.