Skip to main content

The Gap Between Approving a Page and Publishing It

Keep CMS approvals tied to the content people reviewed, with revision records, shared-content checks, and a publishing step that detects later changes.

By VilcorpPublished 6 min read
A displaced cyan glass layer interrupts the aligned green opening in nested graphite contours

An approval can outlive the page it describes

A communications lead reviews a location page on Tuesday. The address, hours, directions, and contact link are correct. They approve it for Thursday's release.

On Wednesday, another editor changes a shared location record. A developer updates the template. Someone replaces the linked PDF. The preview URL still works, and the ticket still says approved. But Thursday's page may contain a combination nobody reviewed.

This hypothetical example matters for healthcare web teams, where public location and service information often comes from several owners. A visitor experiences one page; the publishing team may be managing five separate inputs.

The useful question is simple: what, exactly, did the approval cover?

An approval should identify a specific content revision and the context needed to understand it. Otherwise, it can quietly become permission to publish whatever happens to be current.

Give reviewers a version they can name

A preview URL is a useful place to review. It is not necessarily a fixed record of what appeared there.

A branch preview can keep the same code while reading changing CMS content. A CMS preview may always show the latest draft. Even a revision-specific page can resolve a referenced image or location record to its newest version.

Our guide to preview environments in enterprise web delivery explains how a shared review surface improves delivery. The next step is to make its contents identifiable.

For one publishing candidate, preserve a short record in the existing CMS workflow or release ticket:

  • Content identity: the page ID, revision, language, and intended destination.
  • Included dependencies: the shared records, assets, and links that materially affect the page.
  • Rendering context: the template or application release used for review when it affects the change.
  • Decision: who reviewed which parts, when, and with what unresolved conditions.

Use stable revision or asset identifiers where the system supports them. A screenshot can supplement this record, but it cannot establish which link destination, alternate language, or interactive state was approved.

The mechanism can stay small. A single-page correction may need one revision and one reviewer. A coordinated location launch needs a record of the assembled experience.

Include the content behind the page

Shared content creates useful consistency. It also means that the page's own revision may describe only part of what visitors see.

For the location example, the heading and introduction might belong to the page while opening hours come from a location record, directions from a media asset, and the contact link from a reusable block. Reviewing the page without identifying those dependencies leaves an incomplete publishing decision.

Contentful's release guidance explicitly tells teams to include referenced entries and assets when assembling a release. That is a useful dependency principle; a release grouping alone should not be assumed to freeze every version involved.

Separate the inputs into two practical groups:

  1. Content included in this decision. Pin its reviewed version where supported, or detect changes and request another review before publication.
  2. Content intentionally allowed to change independently. Name its owner and rules. A current service-status notice may need to update without reopening every page that displays it.

Do not claim that an entire page is frozen if some parts remain live. Make that boundary visible to the reviewer.

This is where governance for shared web components supports publishing: the team needs to know which decisions belong to the page and which belong to the reusable component or source record.

Make later edits reopen the right decision

The workflow should preserve the approved candidate while allowing future work to continue.

Drupal's Content Moderation documentation describes keeping a published version live while a separate working copy moves through review. That separation is a useful foundation. The specific approval states, permissions, and dependency checks still need to match the site's configuration.

For the hypothetical location release, imagine that reviewers approve page revision 42 with location revision 17. An editor then creates location revision 18 with different hours.

The publishing step should expose that difference. The owner can approve the new hours with the page, or retain the earlier combination if it is still accurate and the system can publish it safely. An older approved revision should never take priority over information the team now knows is wrong.

Review the changed scope

A revision mismatch should trigger a clear decision, not an unexplained error or an automatic return to the start of a long approval chain.

Show what changed and route it to the relevant owner. A changed phone number needs a content check. A template change that moves the primary contact link may need interaction and accessibility review. An internal scheduling note should not invalidate public copy approval if it cannot affect the published result.

Define those rules deliberately. For small workflows, treating every public-content edit as requiring fresh approval may be the clearest starting point. More selective rules should follow evidence about how the content is actually used.

Check the candidate at publication time

The gap between the final check and the actual write matters too.

Where the platform supports it, publish the identified candidate or use a version check that rejects a write if the content changed after it was read. Checking a revision and then issuing an unrestricted “publish latest” action leaves room for another edit between those steps.

If the CMS cannot enforce that boundary, use its supported locking or release controls and test their limits. A short, scoped editing freeze with a named release owner can be a practical fallback. It should have a clear end and should not stop unrelated publishing.

A useful release sequence is:

  1. Compare the candidate and its included dependencies with the approval record.
  2. Resolve changes with the appropriate reviewers.
  3. Publish through the platform's supported version or concurrency controls.
  4. Record what actually became public, including any partial failures.
  5. Open the public page and verify the critical content and destinations.

For a cached or statically generated site, a successful CMS action may precede the updated public page. Verify the rendered result after the relevant rebuild or cache refresh. In the location example, inspect the hours, address, directions file, and contact link as a visitor would.

Decide how to recover before the release. Reverting a page alone may not restore shared records or replaced assets. Keep the recovery scope as explicit as the publishing scope.

Close the gap with one real publishing workflow

Start with one page type where several people contribute to the result. Follow a change from draft to approval to the public URL.

Can the team identify the reviewed revision? Can it see a dependency change? Does publication preserve that decision? Can an operator verify what visitors received?

Those questions turn approval from a vague status into a usable part of an enterprise web platform. They also help editors keep moving: reviewers can assess a visible change instead of reopening an entire page because nobody knows what happened after sign-off.

The goal is a short, reliable connection between the decision people made and the content visitors receive.

If your publishing process loses that connection, Start a Project to map the revisions, shared content, and release checks around one important page journey.

Keep reading

Explore enterprise web
Newly launched web platform surrounded by concentric monitoring systems
Enterprise webChecklist

The First 72 Hours After an Enterprise Web Launch

A practical operating checklist for the first three days after launch so enterprise web teams can catch business-critical issues before they turn into avoidable fire drills.

5 min read

Put it into practice

Find the next useful platform improvement.

Connect the experience, publishing workflow, and measurement to a delivery plan your team can act on.

Discuss Your Web Platform