Content Decay

Decay costs more than it shows. A page that loses a few clicks a month never sets off an alarm, yet over a year it can lose a large part of what it earned, and the loss is spread across many pages at once. The pages that decay are often the ones that used to perform best, because they were written first and have had the longest to go out of date.

It is also easy to misread. Google’s guide to traffic drops lists the other causes that can look alike on a chart: algorithmic updates, technical issues, security and spam issues, seasonality and changing interests, and site moves^[1]. Each needs a different response, and only decay is cured by work on the page itself. How to tell them apart, page by page, is the subject of the guide to finding and reversing content decay.

A fall in visits from AI assistants has causes of its own. OpenAI says that sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers^[2], and in QueryRev’s study 172 of the 6,294 sites whose robots.txt could be read, 2.7%, block OAI-SearchBot by name^[3]. A new rule in robots.txt can look like decay in a report of AI referral traffic, and it is fixed in a different place.

A new date is not the cure. Google asks site owners whether they are changing the dates of pages to make them seem fresh when the content has not substantially changed^[4]. What reverses decay is a page that serves today’s searcher better than the pages that overtook it.

Content decay is measured page by page, over a long window. In Search Console, the Performance report can show the last 16 months and compare a period year over year, so that each month is read against the same month a year earlier rather than against the month before^[1].

The figure is a page’s clicks in the latest period against the same period a year earlier, written as the share of its own earlier clicks that it lost. Read it beside the page’s impressions and position: Google notes that impressions holding while clicks drop point to the title and snippet, or to a more appealing result from another site^[1].

Keep the review on a fixed schedule, with the same settings each time, and record the day each page was last changed, so that a refresh can later be measured against the pages left alone.

Sample data: an invented site, to show the method.

A payroll software company’s guide to year-end filing ranked near the top for two years. Against the same quarter a year earlier, its clicks fell from 4,200 to 2,900, its impressions from 61,400 to 52,300 and its average position from 2.4 to 4.1. Google Trends showed steady interest in its main queries, and nothing changed on the site’s templates.

The page still names last year’s thresholds, and the two pages that now rank above it answer a question it skips. Updating the figures and adding the missing answer is the decision; the result is read against the same quarter a year earlier, not against the quarter before.