Old company websites can reveal how the organization described its product, customer, market, pricing, partnerships, footprint, and credentials at different points in time. They are especially useful when current deal materials make historical claims that should have left a public trace.
The challenge is not simply finding an old homepage. Domains change, URL structures drift, captures cluster around some dates and disappear around others, and a rendered page may omit assets that were loaded dynamically.
Start with an identity and domain map
List the legal and trading names, current domain, former domains, acquired brands, major product names, and approximate transition dates. Sources for this map can include corporate filings, certificates, releases, email domains in old documents, and redirects observed today.
Keep entity resolution cautious. A similar name or redirected domain does not by itself prove common ownership during the period. Record why each domain was included.
Choose page families before choosing dates
Website comparison is strongest when it compares the same function over time. Build a small page inventory rather than sampling random URLs:
- Homepage and company/about pages for positioning and history
- Product and solution pages for scope, features, and target customer
- Pricing pages for packaging and sales-motion changes
- Customer and partner pages for public relationship references
- Security, trust, and compliance pages for dated control claims
- Location, team, and careers pages for footprint indicators
URLs may change even when the function stays the same. Group `/clients`, `/customers`, and `/case-studies` as possible members of one page family, then retain each original URL as evidence.
Sample change, not just calendar intervals
A capture every January can miss a six-month representation that appeared and disappeared between samples. Begin with broad coverage, then sample more densely around product launches, financings, acquisitions, leadership changes, or other events relevant to the claim.
Deduplicate captures with identical content so a period with heavy crawling does not appear more important merely because it has more records. Preserve failures and gaps; do not silently substitute a nearby page.
Asteron Systems: one homepage, three positions
- 2018 capture
- “Field operations software for regional maintenance teams.”
- 2020 capture
- “The operating system for distributed infrastructure teams.”
- 2023 capture
- “AI coordination for critical infrastructure.”
- Reading
- The captured public positioning broadened by customer scale and later emphasized AI. The pages do not by themselves establish product revenue, capability depth, or why the language changed.
Preserve a citation that survives handoff
For every material observation, record the original URL, archived URL, archive timestamp, exact quotation, page title or classification, and the retrieval date. Where possible, retain a content hash or screenshot linked to the same capture. A screenshot is useful context; searchable text and the archived source are still needed for review.
A capture timestamp establishes that the archive stored a representation by that moment. It does not always establish first publication. Phrase timing accordingly: “This representation first appeared in the captures reviewed on…”
Know what website history cannot tell you
Web archives are selective. Robots rules, crawler failures, authentication, JavaScript rendering, site migrations, and blocked replay can all create gaps. Removed wording may reflect a redesign or disclosure policy, not a change in the underlying fact. Customer logos do not prove active contract status.