Global Trends In The Ephemerality Of Online News

The GDELT Project's Global Difference Graph recrawls each online news article it monitors again after 24 hours and after one week and catalogs any changes made to the article since it was first published, from edits to the article text or title to redirections and even deletions. In all, it has cataloged more than 1.3 billion change records since August 2018. Using this data it is possible to examine some macro-level trends in the ephemerality of online news.

• Deletions: 1.64% of articles were deleted within a week (0.71% within 24 hours). In all, 42.2% of deletions come within the first 24 hours after an article is published.
• Redirections: 2.41% of articles redirected to a different URL within a week (1.87% within 24 hours). In all, 75.2% of redirections come in the first 24 hours.
• Errors + Redirections: Combined, errors and redirections account for 4.03% of URLs within a week (2.59% in 24 hours). In all, 62% of URL-level changes occur within 24 hours.
• Unchanged HTML: 4.51% of articles had completely unchanged HTML within a week (6.42% in 24 hours). In all, 95.9% of HTML changes occur in the first 24 hours.
• Title Change: 6.61% of articles had changes to their title within a week (5.68% within 24 hours). Roughly 83% of title changes come within the first 24 hours.
• Body Text Change: 19.4% of articles had changes to their body text within a week (21% within 24 hours – this number is slightly higher because pages that are deleted or redirected after 24 hours count towards the 24 hour count but not the one-week count). Roughly 82% of body text changes come within the first 24 hours.
• Title + Body Text Change: 15% of articles had changes to their title, body or both within a week (15.1% within 24 hours. Roughly 82% of title and body text changes come within the first 24 hours.

Below you can see each of these indicators plotted at a daily level over time showing that for the most part they have remained remarkably stable over time.