When a web page has changed and nobody saved the old version, the Wayback Machine is usually the first place anyone looks. Often it is the only place. The Internet Archive has been crawling the public web since 1996, and its captures turn up in trademark disputes, contract claims, patent prior-art searches and journalism every day.
It is a remarkable public resource. It is also a record made by someone else, on their schedule, for their purposes. This article covers what that means when the capture has to stand up as evidence.
What a Wayback capture shows
A Wayback capture records that the Internet Archive's crawler requested an address at a particular time and stored what came back. When you view it, the Wayback Machine replays that stored copy, rewriting links so that they point into the archive rather than the live web.
That's valuable evidence, and courts do accept it. Usually it comes in with some support: an affidavit from the Internet Archive, testimony from someone familiar with how the archive works, or an agreement between the parties that the printout is genuine.
Courts don't treat it as beyond question
It is tempting to treat a Wayback capture as self-evidently accurate. At least one US appeals court has said it isn't.
In Weinhoffer v. Davie Shoring, Inc. (5th Cir. 2022), the dispute turned on the terms and conditions of an online auction. The terms limited damages to 20% of the winning bid. The district court took judicial notice of a Wayback Machine copy of those terms, which let them in without further proof. The Fifth Circuit reversed. A private internet archive, it held, falls short of being a source "whose accuracy cannot reasonably be questioned", which is what judicial notice requires.
Other courts have been more accommodating, and the Internet Archive itself asks litigants to try judicial notice or a stipulation with the other side before requesting an affidavit. It provides affidavits for a fee. The point is not that Wayback evidence fails. It's that you should expect to do some work to authenticate it, and that the other side can challenge it.
Where it falls short
You don't choose when it captured
The crawler visits on its own schedule. Popular pages may be captured many times a day. A specific terms page on a small supplier's site may be captured every few months, or never. The capture nearest the date you care about might be weeks either side of it, and the page could have changed in between.
The page may not be there at all
The Internet Archive only collects publicly available pages. It doesn't capture pages behind a login or a form, and some sites are excluded at their owner's request. A site owner who expects a dispute can ask for exclusion, and the history you were relying on disappears from public view.
A replayed page may mix moments
A page is built from many files: the HTML, stylesheets, images, scripts, and data loaded after the page opens. The archive captures each of these separately, often at different times. When it replays a page, it fills each file in with the nearest copy it has, which may be from a different day, or a different year.
Researchers call this temporal incoherence. A 2015 study by Scott Ainsworth, Michael Nelson and Herbert Van de Sompel, Only One Out of Five Archived Web Pages Existed as Presented, found that only about one in five of the archived pages they examined could be shown to have looked the way the archive displayed them. For a page whose meaning is in its text, that may not matter. For a page where a price comes from a script, or the fine print comes from a separate file, it can matter a great deal.
Saving it yourself makes it public
The Internet Archive's Save Page Now lets anyone archive a page on demand, which fixes the timing problem. But the capture is public and searchable by anyone, including the company you're in a dispute with. It is also stored on the same terms as every other Wayback capture, so the site owner can still ask for its removal later.
When to use which
Use the Wayback Machine for the past you didn't capture. If a page changed before you knew it mattered, a public archive may be the only evidence there is. Download the capture, note its exact Wayback URL and timestamp, and talk to your lawyer early about how it will be authenticated.
Keep your own copy of the present. Once you know a page matters, don't rely on a third party to have captured it at the right moment. Save it yourself, privately, with:
- the whole page in a single file that doesn't depend on the live site, captured at one moment;
- the address requested and the address reached;
- a trusted timestamp from an independent authority, obtained at the time of capture;
- a hash, so any later change is detectable;
- a record of who captured it and why.
Our guide on how to preserve a web page as evidence goes through each of these.
Capture on a schedule if the page is likely to change. A weekly capture of a terms page costs almost nothing and means you don't need to hope a crawler passed by at the right time.
Where Archivepanel fits
Archivepanel is the "your own copy" half of this. A real browser saves the page into one self-contained file at a single moment, so there are no parts from other dates. It's private to your workspace, and nobody is told you captured it. DigiCert timestamps the file's fingerprint seconds after capture, and anyone can check that timestamp with OpenSSL, without an account. Scheduled captures keep a dated history of a page or a whole section of a site.
It can't reach back in time. For anything that changed before you started capturing, the Wayback Machine is still the place to look.
Sources
- Weinhoffer v. Davie Shoring, Inc., 23 F.4th 579 (5th Cir. 2022), Justia.
- Internet Archive, Wayback Machine general information, Internet Archive Help Center.
- Scott G. Ainsworth, Michael L. Nelson and Herbert Van de Sompel, Only One Out of Five Archived Web Pages Existed as Presented, ACM Hypertext 2015.