Is E-Discovery Too Expensive?

Recently I’ve had discussions with several lawyers at big firms and at litigation boutiques, all of whom have a clear understanding of their obligations and their clients’ obligations to preserve, review and produce electronic documents, but all of whom seem to be stymied by the apparently uncontrollable, even irrational costs of ediscovery. They have a great deal of difficulty explaining even to sophisticated corporate clients the necessity of paying for electronic discovery, especially since clients number one instruction on document discovery seems to be, “we don’t want to spend any money.”

There’s no question that doing ediscovery properly from beginning to end throughout a contentious matter involving large entities isn’t free and we know that compared to traditional kinds of discovery (which can be done in dribs and drabs over the course of preparation for trial) that using electronic tools and gathering electronically stored information will always involve an upfront cost that is surprising to many and unwanted by all. But in addition to their inability to justify ediscovery to their clients, lawyers often have an irrational fear of ediscovery costs being out of control. These fears derive from some of the early U.S. cases in which parties came to court complaining of potential costs in the millions (for restoring all backup tape sets or searching through all data on a network) and those cases, having been reported, are clearly going to discourage any party from confidently moving forward with an ediscovery plan in hand.

Admittedly there are cases where a lot of money is spent and has to be spent: for example, large-scale investigations or prosecutions (often on the securities side) where cases involve complex transactions, allegations of fraud, a long history of communication amongst numerous parties – in these types of cases it can be very costly to prosecute as well as defend. Another problem with the perceived high cost of ediscovery is that in many cases, lawyers and their clients alike feel that all of the data that has been collected, processed, and to be reviewed, is irrelevant, but it all has to be done anyway, because the Rules say that production has to complete and everything potentially or remotely relevant to any issue in question must be produced. So we go through the exercise to collect and review gigabytes of data reluctantly even with a degree of cynicism and resentment..

To tell one story along these lines, at a recent meeting with senior counsel on a case involving approximately 1.2 terabytes of captured data and a $75,000 proposal, after discussing various aspects of our discovery plan I asked a plain question: “To give us some guidance in terms of how to proceed with the culling of this data, and based on the type of case that this is, what you know already, and the key documents that you’ve already been able to see, is it your gut feeling that these hard drives are likely to contain some helpful or harmful information that you’re anxious to get your hands on? ” Without skipping a beat the answer came back: “No, this stuff is mostly crap.”

That type of situation gives e-discovery a bad name. All of the above can be avoided, all of these issues can be resolved in a way that is reasonable and effective, but certain things have to happen first:

  1. Lawyers have to develop an understanding of the true practical nature of e-discovery and not only gain an understanding of what the potential costs are but what the potential benefits and savings are too. Paying $75,000 to collect, cull, de-dupe, process, and host a million critical documents in a $200 million litigation seems almost free to me. On the other hand, in a nuisance lawsuit, it’s almost criminal.
  2. Lawyers have to understand the importance of an effective plan and an meaningful meet and confer process. Without collaboration with the other side, e-discovery can easily be ten times more costly than it needs to be.
  3. We not only have to read the Sedona Principles, we must demonstrate some courage in applying them. This means that where data sampling and a phased approach is indicated, it should actually be proposed and done. Where we have diligently preserved and collected data from 100 custodians, we should propose to process documents of (say) the 10 key custodians, and review their data first, before moving on (if ever) to the other 90. As part of a collaborative, ongoing process of unfolding discovery, costs can be not only spread out a little bit more over time for the client’s benefit, but costs can be contained at any time in the process, given the rule of diminishing returns.
  4. The approach often taken by law firms to linear review needs to be looked at more closely, and I don’t just mean by incorporating concept search engines, which are obviously being used more commonly (and are adapted into review platforms such as Ringtail and iConect). Not only do these types of tools need to be considered, but other tools such as relevancy ranking, e-mail threading, and near-duplication detection can provide enormous cost savings overall even though they might be associated with an up-front cost.
  5. We know that in most Canadian jurisdictions right now, the Rules are inadequate to deal with the fact that organizations and individuals maintain decentralized and impossibly huge and growing collections of data that are not business records; that are not needed for business purposes; that are not required under document retention or limitations legislation. Lawyers have three options:

    a. Plead the Rules to justify an inordinately expensive discovery (for the other side)
    b. Plead the Rules to justify avoiding e-discovery altogether (for your client)
    c. Use the Rules to fashion a process that works, that is practicable, and whose cost makes some sense.


  1. Iterative approaches to e-discovery were one of the key lessons I took away from LegalTech 2009 last week. That, and the cooperative principles from Sedona. I’m looking forward to seeing if the notions take. This is an interesting entry, and I think pragmatic and sensible.

  2. E-discovery can also be streamlined by judicious use of appropriate search string/s tailor-made to suit the purpose/s of e-discovery. Boolean searches in conjunction with type of documents coded (or NOT coded!) in a particular way can not only help pull out a small subset of relevant docs, but also help stringent QC-ing of large-scale reviews. Platforms like iConect and Stratify offer excellent search capabilities that can be combined with document coding.