Keep three kinds of date separate
The publication or update date belongs to the original page and describes the claimed state of its content. The archive timestamp is embedded in the snapshot address and indicates when the service requested the resource. Your access date records when you inspected the archived copy. None can stand in for another.
A capture may occur months after publication. A page with no displayed date may have many captures. Do not claim that the content was published on the archive date. Use explicit labels such as “archived” and “accessed” so readers can reconstruct the chronology.
A snapshot may be incomplete
Open the exact capture and inspect heading, text, creator, date and the passage you need. Test images, PDFs and links that support your claim. Web archives can retrieve components at different times; an embedded graphic may come from another capture or fail entirely. Forms, video and client-side applications often do not replay fully.
Compare adjacent captures if the layout or text appears broken. Do not automatically choose the oldest or newest calendar point. Choose the snapshot that evidences the state under discussion and record missing components. A limitation note prevents the capture from later being used to support more than it contains.
Take identity from the original resource
The archive is the access route, not normally the author. Read the capture for the responsible organisation, named author, page title and original date. Treat an About or imprint page cautiously if it belongs to a later snapshot. Keep the original URL in the reference because it reveals provenance and site structure.
“Wayback Machine” alone is not a useful citation. Social posts, database records and press releases may need additional fields. Follow the prescribed style while preserving the full archive URL that resolves to the same timestamp.
Lead from the work to the preserved copy
A general structure is: “Author/Organisation (original date), ‘Page title’, original URL, archived by the Internet Archive Wayback Machine on [timestamp], snapshot URL (accessed [your date]).” Adapt order and punctuation to the required style. A note system may place the long archive address in the first citation.
Quotations need a recoverable location. Archived HTML rarely has stable pages, so identify a heading, paragraph or anchor where the style permits. A screenshot can support your audit notes, but it is not a substitute for the citable archive address and may crop context.
Respect exclusions, rights and absent captures
A site can be excluded from replay, and resources may be absent for legal or technical reasons. Do not circumvent access controls. Search a national web archive, library holding or authorised publication. If no capture exists, somebody else's description does not become your own full-text access.
Archiving does not automatically alter copyright, privacy or data-protection obligations. Quote only what is necessary and treat deleted personal material with particular care. This source workflow is general information, not legal advice.
Reopen archived references before submission
Test both snapshot and original URL, note redirects and confirm that the cited passage remains in the selected capture. Archive access can change and records can be removed. For central evidence, an institutional web archive or an additional permitted research deposit may be appropriate. For continually changing pages, use the snapshot and version strategy.
Keep the timestamp as digits in your notes even if the formatted citation displays a readable date. This helps distinguish captures made minutes apart. When a page links to a separate PDF, cite and verify that file as its own object rather than assuming the page snapshot preserved it.
Treat archived downloads as separate sources
If the page links to a report, dataset or PDF, open the archived file and inspect its own title, creator, date and extent. A captured landing page does not prove that the download was preserved, and the file may carry another archive timestamp. Cite the document itself when it supplies the evidence, using the webpage only as context. Record both timestamps and use checksums where appropriate to determine whether files from different captures are identical.