The Maze: AI search has a longer memory than most content calendars. Across 2,000 identical queries per platform, the average cited page was 2.4 years old in ChatGPT and 3.7 years old in Google AI Overviews. Freshness can still help. But the test rejects the lazy idea that every URL needs a cosmetic update to stay visible. Useful pages do not expire because the editorial calendar says they should.
Even the youngest result was old by marketing standards. ChatGPT cited content averaging 891 days, or 2.4 years, while Copilot averaged 1,003 days and Perplexity 1,076. The source test used the same 2,000 queries on each platform. That does not mean recency is irrelevant. It means a page can keep earning citations long after the campaign that launched it has disappeared from the dashboard.
Google's products reached further back. Gemini's average was 1,252 days and Google AI Overviews reached 1,346 days. That is a 455-day spread versus ChatGPT. The source does not explain the mechanism, so it would be reckless to call this an engine preference for old work. Retrieval systems may be rewarding authority, links, topical fit, structure or the simple fact that a durable answer has had longer to accumulate evidence.
Marketers may be overpricing freshness. In a separate NP Digital survey, 91% of 100 content marketers named freshness and recency as a primary driver of AI citability. The observed averages do not disprove that belief. They expose the missing half: newer content can gain an edge while older pages remain highly citable. Freshness is a signal. It is not a shelf-life stamp.
The average hides the distribution. NP Digital does not publish medians, quartiles, query categories, geography or the share of citations under six or twelve months. A few very old sources could lift the mean. Publication dates may also differ from the last substantial update. The safe conclusion is therefore narrow: age alone cannot tell a content team what to refresh, consolidate or retire.
Refresh when the answer decays, not when the date embarrasses you. Prices, product availability, regulation, links and fast-moving comparisons deserve aggressive maintenance. Evergreen guides and durable explainers should be judged by accuracy, performance and citation evidence. Google's own guidance says a new date should follow significant changes, not artificial freshening. A better workflow protects the URL's accumulated trust while adding genuinely new value.
Why it matters: Ecommerce teams sit on years of buying guides, category explainers, product comparisons and policy pages. Treating all of them as perishable creates an expensive treadmill and risks damaging URLs that already carry authority. The smarter portfolio separates content decay from content age: update volatile facts quickly, improve weak pages deliberately and leave useful survivors alone until the evidence says otherwise.
