# Content recency

The visible date on a page is a freshness signal answer engines weight heavily — when it's missing, the page defaults to "stale" regardless of how fresh the content actually is.

By AgentSite · 2 min read · Updated 2026-05-23

Content recency measures whether a page exposes a recent publication or update date — a visible "Last updated" line in the body, a `datePublished` or `dateModified` property in JSON-LD, or both. Answer engines weight recent content heavily. A page without a visible date defaults to "stale" in the engine's ranking, regardless of how fresh the content actually is.

## The two surfaces

The same date should appear in two places:

1.  **Visible in the rendered body.** A "Last updated: 2026-05-24" line near the top or bottom of the page. The agent reads this directly when extracting; a human reader also sees it.
2.  **Encoded in JSON-LD.** [Schema.org](http://Schema.org)'s [`Article`](https://schema.org/Article) type defines `datePublished` (initial publication) and `dateModified` (most recent update) as first-class properties. The agent that parses structured data picks the date up here regardless of where it sits visually on the page.

A page that exposes both gets the strongest freshness signal. A page that exposes only one is fine; both are better. A page that exposes neither defaults to undated, and "undated" reads as "old" to engines that bias toward recency.

## Why engines weight it

The Vercel analysis of AI-crawler traffic for one month in late 2024 is itself a freshness artifact ([Vercel, "The Rise of the AI Crawler," Dec 2024](https://vercel.com/blog/the-rise-of-the-ai-crawler)) — agents that cite it know it's recent, which is part of why it gets cited so heavily. The same logic applies to whatever your page covers: when a user asks about a topic in a domain that changes (regulation, framework versions, AI policy, pricing), the engine reaches for the most recent canonical source it can find.

The Princeton GEO paper measured content-optimization tactics in a controlled experiment and reported visibility lifts of up to 40% across optimization stacks ([Aggarwal et al., KDD 2024](https://arxiv.org/abs/2311.09735)). Recency is among the practitioner-named tactics in that family — fresh content tends to outperform older content on the same topic, all else equal.

## Anti-patterns

Three failures show up reliably:

1.  **Date inflation.** Bumping `dateModified` without changing the content. Major engines detect the no-content-change case and discount it; the signal becomes worthless for that domain over time.
2.  **Hidden dates.** A `dateModified` in the JSON-LD that doesn't match any visible date on the page. Agents that read both surfaces flag the mismatch; readers who land on the page see nothing dating it.
3.  **Inconsistent dates.** Visible body shows `2025-11-12` while JSON-LD shows `2026-03-05`. The engine has to pick one or distrust both. Pick one source of truth and propagate it.

The honest pattern is to update the date _only_ when content meaningfully changes, and keep visible body + schema in sync.

## Where this fits

Content recency is a Layer 4 dimension — content quality at the per-page level. It pairs with [statistics and citations](/statistics-citations) (the same per-paragraph fact backbone benefits from a dated context) and the [direct-answer](/direct-answer) lede (which often includes a year or date as part of the claim). The longer thesis is in [agent readability](/agent-readability).