On this page7 sections
Why it matters for founders and small teams
Your best article from last year loses ground as newer pages appear, and AI engines notice sooner than Google does: ChatGPT’s citations average more than a year newer than Google’s organic results. For a small team that’s leverage as well as risk — updating five pages that already rank is faster than writing five new ones, and a competitor’s stale page is one of the few openings a bigger brand leaves.
Does content freshness affect Google rankings?#
Content freshness affects Google rankings for queries where recent information is expected — news, events, new releases, fast-changing topics — through what Google calls “query deserves freshness” systems, while for stable topics an older page that answers better can still outrank a newer one.
Google’s ranking systems guide describes “various ‘query deserves freshness’ systems designed to show fresher content for queries where it would be expected.” Its examples: someone searching for a movie that’s just been released probably wants recent reviews, and a search for “earthquake” normally returns preparation guides — until an earthquake happens, when news and fresher pages appear.
The idea is old. Google’s 2011 freshness update, which it said affected about 35% of searches, targeted recent events and hot topics, regularly recurring events and subjects that change often. Google weighs freshness query by query, not as a blanket bonus for new pages.
Do AI engines prefer fresh content?#
Most AI engines cite noticeably fresher pages than Google ranks: across about 17 million citations, Ahrefs found AI assistants cited content 25.7% newer than organic results, with ChatGPT leaning hardest and Google AI Overviews showing no freshness premium at all.
| Where the page was cited | Average age of cited pages | Compared with Google organic |
|---|---|---|
| ChatGPT | 958 days | 458 days newer |
| Copilot | 1,056 days | 360 days newer |
| Gemini | 1,118 days | 298 days newer |
| Perplexity | 1,166 days | 250 days newer |
| Google AI Overviews | 1,432 days | 16 days older |
| Google organic results | 1,416 days | — |
Figures from Ahrefs’ July 2025 study, which also found ChatGPT and Perplexity tend to place newer sources earlier in their citations. The engines’ own documentation explains part of the gap:
Perplexity filters stale pages
OfficialIts index stores published and last-updated dates, and prefilters remove “clearly non-responsive or stale content” before ranking begins.
Claude sees each page's age
OfficialEvery Claude web search result reaches the model with a
page_agefield — when the page was last updated — alongside its URL and title.ChatGPT rewards recent updates
ObservedSE Ranking found pages updated within the past three months averaged 6.0 ChatGPT citations against 3.6.
Google applies Search's rules
Our readAI Overviews run on Google’s core ranking systems, so freshness likely matters where query-deserves-freshness systems expect it — consistent with Ahrefs finding no overall premium.
Rankbox benchmark
The Freshness Clock
Five reference points for deciding when a page needs a real update. The first four come from published citation studies; the last is a rule of thumb, labelled as one.
- ≤ 90 days
The ChatGPT refresh window
Pages updated within the past three months averaged 6.0 ChatGPT citations against 3.6 (SE Ranking, 129,000 domains). Keep the pages you most want cited inside this window.
- 458 days
ChatGPT's recency gap
ChatGPT’s cited pages averaged 958 days old, against 1,416 for Google’s organic results (Ahrefs, July 2025). A page that is only as fresh as what ranks is still old by ChatGPT’s standards.
- 250 days
Perplexity's recency gap
Perplexity’s citations averaged 1,166 days old, and its index prefilters stale content before ranking. Give it an honest updated date to read.
- ≈ 0
Google AI Overviews' premium
AI Overviews cited pages 16 days older than organic results on average (Ahrefs). For Google, freshness matters where its query-deserves-freshness systems expect it, not everywhere.
- 12 months
The dated-page rule (rule of thumb)
Any page with a year in its title, a “best” list or prices gets a substantive review at least every 12 months, sooner if its facts change. With the year in 94% of Claude’s sub-queries (Profound), a stale year is a visible mismatch.
How to read it: Sort your pages by last substantive update. Anything that ranks or gets cited and sits past its window goes to the top of the refresh queue; anything past 12 months with no traffic or citations is a consolidation candidate, not a refresh.
Free to use and adapt. If you cite it, link to rankbox.xyz/glossary/content-freshness.
How often should you update content?#
Update content whenever its facts change, and review pages you want cited at least quarterly: SE Ranking found pages updated within the past three months averaged nearly double the ChatGPT citations (6.0 vs 3.6) across 129,000 domains.
What counts is a substantive update — one a returning reader would notice:
- Re-check every number, price and date against its source, and replace anything superseded.
- Answer the questions that appeared since you published — new features, new competitors, new rules — as new sections.
- Cut what’s wrong or dead: outdated screenshots, broken links, advice that no longer applies.
- Update the visible date and
dateModifiedtogether, and only after the content changed. - Keep the URL. Update in place so the page keeps its links and history; a new URL has to earn both again.
Then help engines notice: an accurate lastmod in your XML sitemap, and an IndexNow ping for Bing — Microsoft is one of ChatGPT’s named search providers. Pages that no longer match any live question aren’t refresh candidates at all — they’re content decay to consolidate or retire.
Common mistakes with content freshness#
The most common content freshness mistakes are changing the date without changing the content, putting the current year in titles of pages that weren’t updated, and publishing a new post on a topic an existing page already covers, which splits signals instead of renewing them.
Myth
Bumping the date makes a page fresh.
Reality
Google asks outright whether you change dates “when the content has not substantially changed,” and Perplexity’s index stores both dates. A fake refresh spends trust for nothing.
Myth
Adding this year to the title is enough.
Reality
Claude writes the year into nearly every search, so a dated title helps you match — but only if the page behind it was actually updated that year.
Myth
Freshness matters equally everywhere.
Reality
Ahrefs found Google AI Overviews cite pages about as old as organic results, while ChatGPT’s cited pages ran 458 days newer. Weigh freshness by engine and by query.
Myth
A newer page always beats a better one.
Reality
Freshness is one signal among many. A recent page that doesn’t answer the question still loses — to an older page that does, or to one with more information gain.
Related terms#
- Content & relevanceContent decayThe gradual loss of rankings, traffic and AI citations a page suffers as its information ages, competitors publish better answers or intent shifts — a predictable decline fixed by refreshing, consolidating or retiring the page.Read the entry
- How LLMs answerKnowledge cutoffThe date after which a large language model has no information from its training data, so anything newer — a launch, a price change, a new competitor — reaches its answers only if the model retrieves it from the live web.Read the entry
- Technical SEOXML sitemapA machine-readable file listing a site’s canonical URLs, optionally with last-modified dates, so search engines can discover pages and prioritize recrawling the ones that changed — a discovery aid, not a ranking factor or a guarantee of indexing.Read the entry
- Technical SEOIndexNowAn open protocol that lets a website notify participating search engines, led by Microsoft Bing, the moment a URL is added, updated or deleted, instead of waiting to be recrawled — Google does not use it.Read the entry
- RAGHow LLMs answerRetrieval-augmented generationA technique in which an AI system first retrieves relevant documents from an index and then gives them to a large language model to write its answer, so the response can cite current sources instead of relying only on what the model memorized in training.Read the entry
- Content & relevanceInformation gainThe new information a page adds beyond what other pages on the same topic already say — original data, first-hand experience, a new angle — and it is the leading explanation for why content that rewrites the top results struggles to rank or be cited.Read the entry
Go deeper
Sources
- 1.A guide to Google Search ranking systemsGoogle Search Central · developers.google.com ↗
- 2.Creating helpful, reliable, people-first contentGoogle Search Central · developers.google.com ↗
- 3.Influence your byline dates in Google SearchGoogle Search Central · developers.google.com ↗
- 4.Giving you fresher, more recent search resultsGoogle · googleblog.blogspot.com ↗
- 5.Do AI assistants prefer to cite fresh content?Ahrefs · ahrefs.com ↗
- 6.How to optimize for ChatGPTSE Ranking · seranking.com ↗
- 7.Architecting and evaluating an AI-first search APIPerplexity Research · research.perplexity.ai ↗
- 8.Web search toolClaude Developer Platform · platform.claude.com ↗
- 9.State of AEO 2026Profound (Josh Blyskal) · joshblyskal.com ↗
