Practical guide

Keep article dates accurate when you publish an update

Readers need to know when an article first appeared and what has changed since. Keep those dates aligned across the page, its structured data and the sitemap without making old evidence look new.

A guide published last year can still be useful after a careful revision. The problem starts when a redesign, migration or routine source check makes every date look like today. Readers lose the article's history, and a search engine receives competing clues about when the content was published or substantially changed.

An editor can see the byline, while search engines can also read Article structured data in the page and an XML sitemap listing the site's URLs. Treat these as different views of the same editorial record. Start with what actually happened to the content, then update the fields that describe that event.

Separate the dates before editing the page

The original publication date answers when readers could first access the article. An update date identifies a later revision. Research collection dates say when the underlying observations were gathered. An event date belongs to the event being discussed. None automatically replaces another.

Google recommends a prominent, clearly labeled publication or update date and corresponding structured data. It also advises keeping visible and structured dates consistent, supplying the correct time zone when one is given, and avoiding an event date in place of a page's publication date. These signals help Google interpret dates; they do not guarantee which date appears in search. Google's byline-date guidance

A fictional article timeline showing publication, a later source check without a content change, and a substantial revision, with separate publication and modification fields.

Illustrative editorial history. A source check that leaves the article unchanged is recorded in the review log; it does not become a new publication date.

For a research article, keep the study period near the findings even after a revision. Updating an explanation in November does not turn measurements collected in October into November results. Similarly, moving an existing article to a new domain is not a reason to present its original reporting as newly produced.

Make one change record drive the relevant fields

Before editing, record the existing dates and the reason for the revision. Write a short description such as “Replaced a discontinued setup step and corrected the example.” That is more informative than “Refreshed,” and it gives the next editor a way to check the metadata against the actual work.

Scroll horizontally to read all columns. Use the arrow keys when the table is focused.

Field or labelWhat it describesWhat to check after release
Visible “Published” labelThe article's original publicationIt has not silently become the release date of a later revision.
Visible “Updated” labelA meaningful later revisionIt matches the revision described in the editorial record.
Article datePublishedFirst publicationIt agrees with the visible publication date.
Article dateModifiedThe article's latest modificationIt agrees with the update being represented, with the correct time zone if a time is supplied.
Sitemap lastmodThe page's last significant updateIt reflects a real change to the page, rather than the time the sitemap was rebuilt.

Google's Article documentation defines datePublished and dateModified separately and recommends ISO 8601 values with time-zone information. Keep them in the existing Article object rather than adding a second, contradictory object. Article structured-data reference

For a fictional article first published on October 7 and substantively revised on November 14, the date fields might be:

{
  "datePublished": "2026-10-07T09:00:00-07:00",
  "dateModified": "2026-11-14T09:00:00-08:00"
}

This is a date-field fragment, not complete Article markup or a recommended timestamp for your page. The example uses different offsets because those California dates fall on opposite sides of the daylight-saving transition. Use the actual publication and revision times for your own record, and retain the rest of your valid Article properties.

Keep sitemap changes tied to real page changes

Google says it uses lastmod when the value is consistently and verifiably accurate. Its examples of significant changes include the main content, structured data and links; changing a copyright year alone is not significant. A sitemap generator should therefore preserve unchanged pages' values rather than stamp the entire file with every deployment time. Google's sitemap guidance

A site move or template release can involve real changes that deserve recording. Inspect what changed on each affected page instead of assuming either that every date must change or that none should. Preserve the original publication date when the article itself is being carried over.

Check the published result, including automated fields

Open the live page and inspect its visible dates, rendered Article data and sitemap entry. Check the source HTML too, because a CMS or shared script may add another date that the editor never typed. Resolve contradictory fields at their source instead of inserting another override.

Then read the article as a reader would. If its update label is new but a study or vendor claim is historical, is that distinction still obvious? If the review found no substantive change, keep the source-check date in the review record rather than making the article appear newly published.

Accurate dates describe your content. They are not proof that a crawler fetched the revision or that an AI answer used it. For that separate investigation, use our guide to checking a page version before an AI retest.

Sources and scope

Gist publishes this guide and sells AI visibility software. The recommendations use Google's public documentation checked November 14, 2026 and an original fictional editing example. We did not run a search or AI visibility experiment. Correct dates and valid structured data do not promise indexing, a search-result date, rankings or citations in AI answers.