Publishing an article involves more than placing text on a page. The title, excerpt, date, author, URL, social image and related links all describe the same piece of work. If those details are maintained independently, they can disagree even when the article itself reads well.
Build a publishing workflow that checks the content and its surrounding metadata together. This is especially useful on a static site where articles may be represented by a Markdown file and a separate catalogue entry.
Give every article one identity
Use a unique slug and connect the content file to its catalogue record explicitly. Check for duplicate slugs and missing files before the build reaches production. A fallback page saying that content could not be loaded should not become the published version of a new article.
Keep the title and excerpt specific to the piece. The excerpt should tell a reader what problem the article addresses, rather than repeat a general description of the business.
For a hypothetical engineering blog, an article about retrying imports should not inherit a description about AI strategy because both use the same template. That mismatch affects the archive, feed and social preview as well as the page itself.
Treat dates as editorial facts
Distinguish the intended publication date from the date of a substantive update. Correcting a typo does not necessarily justify presenting an article as newly researched. If a piece is backfilled into an editorial schedule, keep a clear internal record of when the work was prepared and reviewed.
Use the same date values in the visible byline, feed and article markup. A reader should not see one date while the machine-readable data supplies a different unexplained chronology.
Calculate reading time from the actual content using a consistent editorial convention. It is an estimate, but it should not be an arbitrary value copied from another post.
Check the generated page
A Markdown preview cannot establish that the final template works. Inspect headings, lists, links, long titles and the byline in the generated site. Confirm that the template supplies one main heading and that article sections follow a sensible hierarchy.
Check that the social image exists at the URL in the metadata. A generated image route may need its own catalogue entry or pagination logic. A valid URL string is not proof that the asset was built.
Validate the article's structured data and compare it with the visible page. Watch for literal template placeholders, duplicated authors or hard-coded reading metrics that no longer describe the content.
Make discovery part of completion
Confirm that the article appears in the archive, category filtering, RSS feed and sitemap where intended. Review the order so the newest material is easy to find. Link to relevant existing articles and service pages where those links help the reader continue.
Choose featured status deliberately. Automatically featuring every new post can make a carefully selected reading list meaningless, while never reviewing old selections can leave the homepage describing an outdated focus.
A compact publishing checklist is usually enough: complete content, accurate metadata, working routes and assets, sensible links, and a rendered-page review. Our engineering approach applies the same principle to content delivery as to software features: verify the thing the reader receives, not only the files used to build it.