Datos estructurados de artículos: comprueba que cuentan la misma historia

Collect visible facts before writing code

Record the edition URL, title, real editorial byline, first publication and latest substantive update. Add the image representing the article where appropriate. Identify the source of each value: an editorial record, published text or a resource you can open. This short list prevents two systems from inventing different versions of one page.

Google documents Article as information that helps it understand an article. In our fictional exercise, a magazine publishes a notebook-care guide. Its author is the organization that actually edits it. There is no need to invent a person to populate authorship.

Use a small fragment you can compare

This is an original abbreviated JSON-LD example. Its addresses use an illustrative domain and its dates are fictional; replace them with the actual page’s data. The fragment is displayed as readable code and does not add a second editorial identity to this guide.

{
  "@context": "https://schema.org",
  "@type": "BlogPosting",
  "headline": "How to Preserve a Notebook",
  "url": "https://journal.example/en/notebooks/",
  "inLanguage": "en",
  "author": {
    "@type": "Organization",
    "name": "Example Magazine"
  },
  "datePublished": "2026-10-01T09:00:00-06:00",
  "dateModified": "2026-10-12T11:30:00-06:00"
}

Compare each value with the page. If a CMS already generates another block for the same article, review their relationship before pasting a contradictory second version. Maintain one editorial data source that can feed both visible HTML and markup.

Describe a recognizable reading path

BreadcrumbList orders its items by position. A possible example path is Magazine → Resources → Notebook Care. Check that the labels help someone recognize the guide’s location and that intermediate destinations exist. The hierarchy can express editorial organization even when the URL text has a different shape.

Our example has positions 1, 2 and 3, with names and URLs for each step. When implementing it, consult the current properties in the Breadcrumb documentation. Our bilingual internal-links guide helps decide which readings connect.

Check every edition and its dates

A Chinese edition needs its Chinese title and address; the English edition needs its own. Also review the declared language and that edition’s breadcrumb destinations. Keep this description distinct from the tags linking translations, explained in our hreflang guide.

Google documents datePublished and dateModified as different moments. Publishing another article or rebuilding the site does not update the whole archive. Preserve first publication and change the modification date when that edition receives a substantive revision. Record timezone information too.

Keep three separate check results

• Syntax: the block is interpreted correctly as JSON-LD.
• Correspondence: values, entities and dates describe visible content.
• Publication: the page and necessary links and images are accessible at their intended URLs.

Google recommends testing and URL-inspection tools within publication. Record the tool, date, URL or code checked and observation. Testing pasted code and testing a published URL cover different things. Our technical foundations guide completes the access review.

Repeat the review when article identity changes

Check markup after changing a byline, removing an image, moving a section or correcting dates. A simple table can record field, visible value, declared value and open decision. If they differ, fix the source generating the incorrect value before spreading the change across other pages.

Google’s policies ask data to represent page content and explain that markup alone does not ensure a special presentation. The useful result is a coherent, inspectable description. Observed search changes can be recorded afterward with their date and evidence.

写代码前先收集可见事实

记录版本网址、标题、真实编辑署名、首次发布时间和最近一次实质更新。适用时加入代表文章的图片。为每个值注明来源：编辑记录、已发布文字或能够打开的资源。这样可以避免两个系统为同一页面生成不同事实。

Google 将 Article 说明为帮助理解文章的信息。在虚构练习中，一本杂志发表手账保养指南，作者就是实际负责编辑的组织，无须为填写作者字段而虚构个人。

采用容易对照的小片段

以下是原创的简化 JSON-LD 示例。网址使用示例域名，日期也属虚构，实际使用时应替换为真实页面资料。它作为阅读代码展示，不会给本指南加入第二个编辑身份。

{
  "@context": "https://schema.org",
  "@type": "BlogPosting",
  "headline": "如何保存一本手账",
  "url": "https://journal.example/zh/notebooks/",
  "inLanguage": "zh-Hans",
  "author": {
    "@type": "Organization",
    "name": "示例杂志"
  },
  "datePublished": "2026-10-01T09:00:00-06:00",
  "dateModified": "2026-10-12T11:30:00-06:00"
}

逐项与页面比较。如果 CMS 已为同一文章生成另一段数据，应先检查两者关系，再决定是否添加第二份。保留能同时生成可见 HTML 与标记的编辑数据来源，避免互相矛盾。

描述读者认得出的阅读路径

BreadcrumbList 按 position 排列项目。示例路径可以是“杂志 → 资源 → 手账保养”。检查标签是否帮助辨认指南位置，以及中间目标是否存在。层级可以表达编辑组织方式，而不必完全复制网址文字。

示例包含位置 1、2、3，各自拥有名称与网址。实现时应查看Breadcrumb 现行文档。本站双语内部链接指南帮助判断不同阅读内容如何连接。

逐个检查版本与日期

中文版需要中文标题及对应网址，英文版也要使用自己的信息。同时检查声明语言与该版本的面包屑目标。这项描述与连接翻译版本的标签分工不同，后者见本站hreflang 指南。

Google 分别说明 datePublished 和 dateModified。发表另一文章或重新构建网站，不代表整个档案都更新了。保留首次发布时间，在该版本有实质修订时才更新修改日期，并记录时区。

分别保留三项检查结果

• 语法：数据块能够作为 JSON-LD 正确解析。
• 一致性：字段、实体与日期描述可见内容。
• 发布：页面及所需链接、图片在预期网址可以访问。

Google 在发布流程中建议使用测试与网址检查工具。记录工具、日期、检查的网址或代码及观察。粘贴代码测试与检查已发布网址的范围不同。本站技术基础指南补充访问检查。

文章身份信息变化时重新审核

署名更改、图片移除、栏目移动或日期更正后，都要检查标记。可以用简表记录字段、可见值、声明值与待决事项。如果不同，先修正产生错误值的来源，再推广到其他页面。

Google 的政策要求数据代表页面内容，并说明仅有标记不保证特殊展示。审核的实际成果是清楚且能够核查的描述。搜索中出现的变化则在之后连同日期与证据另外记录。