A fellow professor of e-discovery started my morning with a question. He wrote, “In companies using Google business, internal email ‘attachments’ are often linked with a URL to documents on a Google drive rather than actually ‘attached’ to the email…. Can the producing party legally refuse to produce the document as an attachment to the email showing the family? Other links in the email to, for example, a website need not be produced.“

I replied that I didn’t have the definitive answer, but I had a considered opinion. First, I challenged the assertion, “Other links in the email to, for example, a website need not be produced.”

Typically, the link must be produced because it’s part of the relevant and responsive, non-privileged message.  But, the link is just a pointer, a path to an item, and the discoverability of the link’s target hinges upon whether the (non-privileged) target is responsive AND within the care, custody or subject to the control of the producing party.

For the hypothetical case, I assume that the transmittal is deemed relevant and the linked targets are either relevant by virtue of their being linked to the transmittal or independently relevant and responsive.  I also assume that the linked target remains in the care, custody or subject to the control of the producing party because it has a legal and practical right of access to the repository where the linked target resides; that is, the producing party CAN access the linked item, even if they would rather not retrieve the relevant, responsive and non-privileged content to which the custodian has linked the transmittal.

If the link is not broken and the custodian of the message could click the link and access the linked target, where is the undue burden and cost?  Certainly I well know that collection is often delegated to persons other than the custodian, but shouldn’t we measure undue burden and cost from the standpoint of the custodian under the legal duty to preserve and produce, NOT from the perspective of a proxy engaged to collect, but lacking the custodian’s ability to collect, the linked target? Viewed in this light, I don’t see where the law excuses the producing party from collecting and producing the linked target

The difficulty in collection cited results from the producing party contracting to delegate storage to a third-party Cloud Provider, linking to information relegated to the Cloud Provider’s custody.  In certain respects, it’s like the defendant in Columbia Pictures v. Bunnell, who put a contractor (Panther) in control of the IP addresses of the persons trading pirated movies via the defendant’s platform.  Just because you enlist someone to keep your data on your behalf doesn’t defeat your ultimate right of control or your duty of production. 

Having addressed duty, let’s turn to feasibility, which is really what the fight’s about. 

Two key issues I see are: 

1. What if the link is broken by the passage of time?  If the target cannot be collected after reasonable efforts to do so, then it may be infeasible to effect the collection via the link or via pairing the link address to the addresses of the contents of the repository (as by using the Download-Link-Generator tool I highlighted here).  If there is simply no way a link created before a legal hold duty attached can be tied to its target, then you can’t do it, and the Court can’t order the impossible.  But, you can’t just label something “impossible” because you’d rather not do it.  You must make reasonable efforts and you must prove infeasibility. Courts should look askance at claims of infeasibility asserted by producing parties who have created the very situations that make it harder to obtain discovery.

2. What if the content of the target has changed since the time it was linked?  This is where the debate gets stickier, and where I have little empathy for a producing party who expects to be excused from production on the basis that it altered the evidence.  If the evidence has changed to the point where its relevance is in question because it may have been materially changed after linking, then the burden to prove the material change (and diminished relevance) falls on the producing party, not the requesting party.  Else, you take your evidence as you find it, and you produce it as it exists at the time of preservation and collection.  The possibility that it changed goes to its admissibility and weight, not to its discoverability.

I hope you agree my analysis is sound. To paraphrase Abraham Lincoln, you cannot murder your parents and then seek leniency because you’re an orphan. The problem is solvable, but it will be resolved only when Courts supply the necessary incentive by ordering collection and production. Integrating a hash value of the target within the link might go a long way to curing this Humpty-Dumpty dilemma; then, the target can be readily identified AND proven to be in the same state as when the link was created.

While we are at it, embedded links should be addressed from the standpoint of security and ethics. If a producing party supplies a message or document with a live link and opposing counsel’s clicks on the link exposing information not meant to be produced, whose head should roll there? If a party produces a live link in an email, is it reasonable to assume that the target was delivered, too? To my mind, the link is fair game, just as the attachment would be had it been embedded in the message. Electronic delivery is delivery. We have rules governing inadvertent production of privileged content, but not for the scenario described.