Find Decaying Articles from a GA4 Pages and Screens Export

How to export monthly page traffic from GA4

Google Analytics 4's built-in comparison date picker only compares two flat windows (such as last 30 days vs. previous 30 days), which misses articles that peaked six months ago and declined steadily each month.

1. In Google Analytics 4, navigate to Reports > Engagement > Pages and screens.
2. Set the date range to the last 12 months. Set the primary dimension to Page path and screen class, then click + to add Year month (formatted YYYYMM) as a secondary dimension.
3. Click Share this report > Download File > Download CSV (exports up to 100,000 rows as of October 2026).

Formulas to detect and rank decaying content

1. Peak Monthly Views per URL (FIXED LOD)

First, lock in the highest monthly view count each URL ever achieved inside your 12-month export window.

xmetric / Tableau formula ([Peak Monthly Views]):

{ FIXED [Page path and screen class] : MAX([Views]) }

SQL equivalent:

WITH monthly_page_views AS (
  SELECT
    "Page path and screen class" AS page_path,
    "Year month" AS year_month,
    SUM("Views") AS views,
    SUM("Key events") AS key_events
  FROM ga4_pages_and_screens
  GROUP BY 1, 2
),
peak_per_page AS (
  SELECT
    page_path,
     year_month,
    views,
    key_events,
    MAX(views) OVER (PARTITION BY page_path) AS peak_monthly_views,
    FIRST_VALUE(year_month) OVER (PARTITION BY page_path ORDER BY views DESC) AS peak_month,
    MAX(year_month) OVER () AS latest_month
  FROM monthly_page_views
)
SELECT
  page_path,
  peak_month,
  peak_monthly_views,
  views AS latest_month_views,
  (peak_monthly_views - views) AS views_lost_from_peak,
  (views - peak_monthly_views) * 1.0 / NULLIF(peak_monthly_views, 0) AS decay_pct
FROM peak_per_page
WHERE year_month = latest_month
  AND peak_month < latest_month
  AND peak_monthly_views >= 500
ORDER BY views_lost_from_peak DESC;

2. Views Lost Since Peak (Refresh Priority Score)

For example (sample numbers), a page that dropped from 10,000 views to 6,000 views (-40%, 4,000 views lost) is worth refreshing before a page that dropped from 200 views to 40 views (-80%, 160 views lost). Sort your refresh backlog by absolute views or key events lost from peak, not percentage drop alone.

xmetric / Tableau formula:

IF [Year month] = { FIXED : MAX([Year month]) }
THEN [Peak Monthly Views] - [Views]
END

3. Months Since Peak (Table Calculation / LOD)

Differentiate a seasonal spike from a multi-month decay trend by checking whether traffic has declined for at least 3 consecutive months after its peak month.

xmetric / Tableau formula ([Peak Month]):

{ FIXED [Page path and screen class] :
  MAX(IF [Views] = [Peak Monthly Views] THEN [Year month] END)
}

Limits of GA4 exports for content audits

• 100,000-row export ceiling in the GA4 UI: Because Page path and screen class crossed with 12 Year month values multiplies row count by 12, sites with more than ~8,300 active URLs will hit the 100,000-row UI CSV export limit (as of October 2026). Filter the GA4 report to your content subfolder (such as Page path and screen class contains /blog/) before exporting, or export from GA4 Explorations (TSV/CSV) or from BigQuery as CSV or Parquet (the xmetric app imports both).
• All Traffic vs. Organic Search segment: A viral social post will look like "content decay" the following month unless you filter the GA4 report to Session primary channel group = Organic Search before downloading the CSV.