Google deployed a new artificial intelligence-powered algorithm called Supersession-Decay Filtering (SDF) in its Discover feed to proactively identify and remove stale content, the company announced.
The system, operational for two years, aims to enhance content freshness and user engagement for hundreds of millions of daily Google Discover users globally.
SDF employs two distinct filters: one for “supersession,” which targets content replaced by newer information, and another for “relevance decay,” addressing content that loses its usefulness over time, according to Google.
This proactive filtering prunes outdated content before it reaches the more computation-heavy ranking stage, which continues to rely on engagement and trend-aware systems.
Traditional methods for identifying content staleness, such as publication dates and reactive user engagement metrics, were deemed insufficient by Google.
Internal testing demonstrated that SDF reduced complaints about stale content by 54.9 percent. The algorithm also improved overall user engagement and content diversity, while simultaneously reducing computing costs for the platform.
Publishers and search engine optimization (SEO) professionals will need to adjust their content creation and interlinking strategies to account for SDF’s proactive detection of content staleness, Google indicated.
Source: Search Engine Journal
