Choose What an Editorial Archive Must Preserve

TL;DR

• Decide the archive’s purpose by answering what a future editor needs to know, and explicitly separate working material, accepted revisions and published records so retention is proportional and accountable.
• Define which records to keep by listing artifacts that let someone reconstruct a meaningful publishing decision, including accepted revisions, publication references, sources, asset references and change notes.
• Make retrieval and review part of the policy: test finding a supporting version with a real question, assign an owner to review and document exceptions so archives remain usable and tidy.

Decide what the archive is for

A content archive policy should begin with the questions a future editor needs to answer. What did the article say? Which version was approved? Why was it changed or retired? Which source supported an important claim?

Those questions do not necessarily require retaining every temporary file forever. They do require a deliberate distinction between working material, accepted revisions and published records.

Treat this as an editorial operating policy. If the organization has separate legal, contractual or security retention requirements, have the responsible owner supply those requirements rather than asking the editorial team to invent them.

Define the records worth preserving

List the artifacts needed to reconstruct a meaningful publishing decision. A final body file alone may omit the image, metadata or source context that made the article understandable.

Keep the record proportional. A minor punctuation edit does not need the same narrative as a corrected product recommendation, but both should have an identifiable revision.

Separate archive state from public state

An archived editorial record may correspond to a live page, a redirected page or a removed page. Do not use one label to imply all three.

Record the public disposition explicitly. If a page was retired, state whether a replacement exists and which team owns the live-site action. A file moved into an archive folder does not itself change the website.

Our guide to updating old posts can help decide whether content should be revised or retired. The archive policy determines what evidence remains after that decision.

Keep sensitive working material separate

Research notes can contain private contact details, unpublished product information or comments that are not appropriate for broad access. An archive should not become an uncontrolled collection merely because the article itself was public.

Classify the material and use the access and retention rules supplied by the responsible organizational owners. Preserve only what the editorial purpose requires within those constraints.

For an illustrative example, an approved factual summary may be useful to retain while a raw interview recording has a different access requirement. Link the records through an identifier without assuming they should live in the same shared folder.

Make retrieval part of the policy

A record is not useful if a new editor cannot locate it. Use stable article identifiers, searchable titles and a manifest that connects revisions, sources and destinations.

Test retrieval with a real question: ask someone to identify the version that supported a published claim six months earlier. Note which missing fields or inaccessible files prevent them from answering.

The content approval workflow should produce the acceptance reference the archive relies on. An archive cannot reconstruct an approval that was never recorded.

Review the archive deliberately

Assign an owner to review outdated or redundant material according to the agreed policy. Check references before removing an asset that may still be used by a live article.

Document exceptions and disposal decisions where required by the organization's rules. Avoid informal cleanup based only on a folder's age or a filename that looks obsolete.

A useful editorial archive preserves enough context to support maintenance and accountability. It remains distinct from the publishing queue, has clear access boundaries and can answer practical questions without forcing the next editor to guess.