Definition
A retained source is the material kept from a collection event so that a finding can be reviewed later. It may be a saved page or document with its source location and collection context. Retention matters because a live URL can change, move, or disappear after a researcher relies on it.
Keeping a page preserves what was read; it does not establish that every statement on the page is true. Company marketing, regulatory records, directories, and archived material answer different questions. A retained source also needs the correct company identity and date context. A historical page can support a historical statement while being insufficient for a claim about current services or ownership. Those limits belong with the finding.
In origination work
Save the supporting material at collection time and connect the relevant passage to the typed answer. Record when it was collected and, where known, when the underlying statement was published. Docket's Audit works from retained material with fetch tools denied, so gaps in that material cannot be repaired by silently consulting a fresh page during review. If support is missing or inadequate, the answer should remain open or be removed from accepted findings until a supported collection is available.
Why keep a copy
A retained source is a stored copy of the page or document from which a finding was taken, kept as retrieved, with the date and address. The reason is that the web changes. A study by the Pew Research Center found that a substantial share of web pages that existed in earlier years were no longer accessible when checked, and legal and academic practice has developed tools such as Perma.cc to preserve cited pages. If a finding depends on a page, the page must be kept.
The Internet Archive's Wayback Machine offers another route: it stores captures of public pages and provides interfaces to find captures by address and date. A capture in the archive can corroborate what a page said at a time, but it is not under your control and may not contain the capture you need. A team that needs reliability keeps its own copy and uses the archive as a second source.
What to store
Store the response as received, not only the rendered text: the HTML or PDF, the retrieval time, the address requested and the final address after redirects, and the response headers where relevant. Extract text from the stored copy and keep the extraction method with the record. Take a screenshot where the page is dynamic, since a copy of the HTML may not show what a reader saw.
Respect the publisher's rules. The robots exclusion protocol, now standardised, tells automated clients which paths a site asks them not to fetch, and site terms may restrict collection or reuse. Courts have considered whether accessing public pages violates computer-misuse law, with outcomes that turn on authorisation, and platform terms of service are a separate matter. Retaining a copy for evidence does not override any of these.
Retained sources in review
The test of a retained source is whether a reviewer who has never seen the original can confirm the finding from the copy alone. Include in review a check that the quoted passage appears verbatim in the stored copy. If it does not, the finding is struck. This is a mechanical check that catches paraphrase drift and mistaken extraction.
Plan retention. Decide how long copies are kept, who may access them, and how personal data in them is handled. Public pages can contain personal information, and data-protection law may apply to storing it. Keep what the research needs, protect it, and delete what you no longer need.
Related terms
- Provenance — The record of where information came from, how it was produced, and which activities contributed to it.
- Evidence trail — A reviewable chain connecting a research conclusion to its supporting material, interpretation, and decisions.
- Grounding — Basing generated research answers on supplied evidence while still checking that the evidence supports each answer.
Further reading
- Collection and audit as separate stages
- Evidence standards
- Evidence trail
- Wayback Machine
- Company websites
- Trust and data practices
- Article: how often to refresh target research
Sources and further reading
- When Online Content Disappears, Pew Research Center (2024). Measures how much of the web cited a few years ago is already gone. The argument for retaining the page, not the link.
- Perma.cc and the problem of link rot in citations, Harvard Law School Library Innovation Lab. Why legal and academic citation moved to archived copies rather than live URLs.
- Wayback Machine, Internet Archive. The public record of what a page said on a date, when you did not retain it yourself.
- Wayback Machine APIs, Internet Archive. How to look up archived captures of a page programmatically.
- RFC 9309 — Robots Exclusion Protocol, IETF. The standardised version of robots.txt, published in 2022 after twenty-eight years as a convention.
- hiQ Labs, Inc. v. LinkedIn Corp., U.S. Court of Appeals for the Ninth Circuit (2022). On whether scraping a public profile violates the Computer Fraud and Abuse Act. The answer turns on authorisation, not on whether you used a browser.
To apply these distinctions to your own acquisition criteria, discuss your research brief or inspect the demo platform.