Definition
Content freshness is how recently a page was published or substantially updated, and whether its information is still current. Search engines have used freshness as a signal for years, mainly for queries where recent information matters, such as news, prices and product releases. AI answer engines appear to weigh it more heavily, especially ChatGPT.
Several studies point the same way:
- A 2025 Ahrefs study of about 17 million citations found that pages cited by AI assistants were on average 25.7% newer than pages in Google's organic results (1,064 days old versus 1,432). ChatGPT cited the newest pages, while Google's AI Overviews cited slightly older pages than organic results (Ahrefs).
- In an October 2025 analysis of ChatGPT's 1,000 most-cited pages, Ahrefs found that 76.4% of the 564 pages with a detectable update date had been updated in the previous 30 days. Wikipedia made up 53% of the sample, which inflates the figure.
- A 2025 academic paper tested seven language models used as rerankers and found that adding a recent 'published on' date to otherwise identical passages moved them up the ranking, in some cases by as many as 95 positions (arXiv).
The last finding is a warning as much as an opportunity. Models can be swayed by a date alone, but Google treats fake freshness as a negative: its helpful content guidance lists 'changing the date of pages to make them seem fresh when the content has not substantially changed' as a warning sign. Freshness also matters most where facts change. Ahrefs noted that the average AI-cited page was still almost three years old, so durable reference content keeps being cited as long as it stays accurate.
Systems read freshness from several places: visible dates on the page, structured data such as datePublished and dateModified in schema markup, lastmod values in XML sitemaps, and the content itself, such as years, version numbers and prices. Google API documentation that leaked in 2024 referred to dates taken from the byline, the URL and the page text. Keep these consistent; a 2026 'last updated' label above 2023 prices sends mixed signals.
A practical freshness program sorts pages by how fast their facts change. Pricing pages, product comparisons, best-of lists and anything with version numbers need checking monthly or whenever the market moves. Guides and how-tos need a review a few times a year. Definitions and evergreen explainers need an accuracy check when something in the field changes. Promptwatch's ChatGPT content half-life data from October 2026 is a useful guide to priorities: ChatGPT kept citing most landing pages and docs for months, product pages and how-tos lost citations faster, and half of news pages dropped out within about five weeks. Pair the schedule with monitoring so that pages losing citations move to the front of the queue; content decay covers detection.
Examples of Content Freshness
- A CRM review site checks pricing on its comparison page every month and keeps a visible changelog of each change. Its page stays among ChatGPT's cited sources for CRM pricing prompts, while a competitor's page, last updated a year earlier, drops out.
- A tax software company updates its retirement contribution limits page on the day the IRS announces new limits, so AI answers quoting the new figures can cite it.
- A content team bulk-updates the dateModified field on 500 posts without editing them. Search traffic does not improve, and a later audit flags the mismatch between the new dates and years-old screenshots.
- A B2B publisher sets review cycles by page type (pricing monthly, guides quarterly, glossary yearly) and moves any page that loses AI citations to the front of the queue.
