Content decay

Content decay is a loss of traffic on a page over time. MeasureScale measures it by comparing a current period with an earlier period of the same length. MeasureScale uses this method to list the pages that lost traffic. It also uses it to classify the whole site over several time spans, and to classify each page.

Query and page trends use the same classifier to label their click trend.

How to read the answer

• Each page gets a status such as growing, stable, moderate, severe, new or lost.
• Severe means the page kept half or less of its traffic from the previous period.
• A lost page had traffic before and none now. That is a real decline, not missing data.
• For the whole site, read all the time horizons together. One bad month is not the same as a year of decline.
• Next, ask your assistant to list the decaying pages, largest loss first.

What it answers

• Which pages lost at least a given share of their clicks or sessions compared with the previous period?
• Is the site’s traffic growing, stable or declining over 1, 3, 6 and 12 months?
• How many pages are growing, stable, declining, new or lost?

Inputs

You can ask about either of two sources. Search Console is the default.

• Search Console measures clicks. It reads the top 1,000 pages by clicks in each window.
• GA4 measures sessions. It reads GA4 landing pages, ordered by sessions.

GA4 needs a GA4 property linked to the Search Console property. Without one, the result says so. It links to the MeasureScale page where you link a property.

A settled day is a day whose data Google has finished processing. With no dates, MeasureScale uses the last 28 settled days as the current window. Data windows and freshness explains the windows in detail.

How it works

• MeasureScale compares a page’s traffic in the period you ask about with the period of the same length just before it.
• The page list shows the pages whose drop reaches the threshold, largest loss first. The default threshold is a 20% drop. You can ask for a smaller or larger one.
• A page needs at least 50 clicks or sessions in the earlier period to be listed. You can ask for a different minimum.
• The site view compares the site’s recent traffic with earlier blocks at each horizon.
• The default horizons are 1, 3, 6 and 12 months. You can ask for other horizons, from 28 to 459 days.
• One classifier turns each comparison into a status.

The decay classifier

The ratio is the current total as a share of the prior total. The classifier checks the conditions in this order:

Each boundary value belongs to the more severe status. A page that exactly halves its traffic is severe. A ratio of exactly 0.85 is moderate. These boundaries are fixed.

A page that had traffic in the prior window and none in the current window is lost. It is a decline, not missing data.

The site’s overall status is the worst of lost, severe and moderate when any horizon has one. Otherwise, it is the most common status among the horizons with data.

Limits

• Lookback. Your plan sets the earliest date MeasureScale can read. If the prior window starts before that date, there is nothing to compare. The page list is then empty, and the result says the data was too short.
• Short history for a horizon. A horizon whose prior block starts before the earliest readable date reports no numbers. MeasureScale never counts days outside the data as 0.
• The 459-day maximum. A 459-day horizon keeps the oldest prior block inside 16 months of Search Console history.
• Row caps. A page missing from a pull that reached its row cap ranks below the cap. It did not necessarily stop earning traffic. MeasureScale therefore leaves such a page out instead of calling it lost or new. The result counts the pages left out.
• Days with no rows. Inside the data, a day with no rows counts as 0. Search Console sends no row for a day without traffic.
• Site-wide view. The horizons describe the whole site. Use the page list for individual pages.
• GA4 sampling. For GA4, the result says whether GA4 sampled or thresholded its reports.