Hot on the heels of its first endorsement of the use of predictive coding in the widely publicised Pyrrho decision in February 2016, the English court has recently given judgment ordering the use of predictive coding in circumstances where its use was opposed by one party.

This recent judgment, made in David Brown v BCA Trading Limited & Ors, is likely to further advance the discussion around the use of predictive coding – one of a number of forms of technology assisted review (TAR) – which has been a litigation hot topic of 2016 following the Pyrrho decision.

Background

The authors' previous article on the Pyrrho decision summarised the reasons for the approval of predictive coding in that case, and also commented more widely upon the principles behind predictive coding technology and the disclosure obligations placed upon parties engaged in English civil litigation (please see "A (bright) green light for predictive coding in disclosure").These matters are not recited in this article.

Since the Pyrrho decision, TAR and the effective management of electronic disclosure in English litigation has become a prominent topic in the legal industry.One overriding message permeating much of the discussions is that the results achieved by utilising predictive coding are likely to be as good, if not better, than by utilising a standard "linear review" of documents by humans, whilst also achieving significant costs and time savings.That stands to reason: in principle, predictive coding allows decisions on relevance to be led by a senior case expert to a much greater extent than a traditional linear review, where many decisions on relevance are ordinarily left to multiple junior lawyers' or paralegals' subjective, independent assessments (often under time pressure).Lawyers should therefore not be wary of predictive coding or other forms of TAR; they should be encouraged and excited by the potential benefits.

Facts

BCA Trading is an unfair prejudice petition under section 994 of the Companies Act 2006 under which the petitioner, Mr Brown seeks a buyout of his minority shareholding in BCA Trading Limited (BCA).

In common with the position in Pyrrho, the majority of the documents likely to be relevant to the dispute are held by one party (BCA), with the other parties including Mr Brown having comparatively few relevant documents.As such, it is perhaps unsurprising that Mr Brown was keen to ensure that he obtained the widest possible disclosure from BCA to help him prove his case.On the other hand, from BCA's perspective there was clearly a significant imbalance in the likely time and costs burden of conducting disclosure which it would want to keep under control by employing appropriate techniques to limit the scope of and/or assist the process.

In light of the above, BCA had proposed utilising predictive coding to ease the burden of conducting its disclosure.It estimated that the costs of conducting disclosure using predictive coding would be in the region of £132,000, with such costs increasing significantly to between £250,000 and £338,000 if a more traditional methodology using key words and manual review was used.Notwithstanding the apparently clear costs benefits, Mr Brown contested the use of predictive coding.

Decision

The decision was made at a case management conference at which a number of other issues were also considered.The part of the judgment relating to predictive coding endorses predictive coding in clear, concise and unequivocal terms.The court placed significant weight upon the anticipated costs savings, and the fact that the disclosing party (BCA) wished to use predictive coding.The court noted that there was no factual or expert evidence before it to contradict the assertions made by BCA.

The court also cited the 10 reasons for endorsing predictive coding recorded in the Pyrrho decision.It noted noting that, whilst one was neutral, all but one of the others applied in this case.(The only reason that did not apply was that the parties had not reached agreement on the use of predictive coding, as they had in Pyrrho.)Whilst the court did not explain in the judgment the basis on which Mr Brown sought to resist the use of predictive coding, it held that there were "no factors of any weight" pointing against use of predictive coding.

In addition, the court commented upon the importance of ensuring that the issues by reference to which the relevance of documents was to be determined for the purposes of disclosure were, to the extent possible, narrowed and agreed before the disclosure process was conducted.The court noted that the statements of case in the proceedings were cast on a wide basis, and that:

"…experience shows that issues will narrow significantly by the time the trial is reached.This can mean that what may have appeared to be necessary disclosure based upon the statements of case at this stage, will turn out to have been unnecessary and indeed to a large degree irrelevant to the way the case will be heard at trial.It can mean that costs will have been incurred which need not have been incurred both during disclosure and when complying with subsequent directions concerning evidence."

In light of this, the judgment confirmed that the court had, as a first stage, proposed a process whereby the parties should seek to identify and narrow down issues by way of schedule before turning to disclosure.Whilst the court did not elaborate on the detail of this proposal, it stressed that it was not an attempt to narrow the disclosure to be given from standard disclosure to issue-based disclosure.Instead, the intention appears to be to encourage a sensible dialogue between the parties at an early stage to make the (standard) disclosure process as targeted, and therefore proportionate, as possible.

Comment

The decision is a helpful continuation of the pro-TAR momentum initiated by the Pyrrho judgment.The BCA Trading judgment comments upon predictive coding in an extremely positive manner, and there are few (if any) comments casting doubt upon the effectiveness or use of the technology.This may signal a continuing shift by the English courts towards a position where predictive coding (or at least some form of TAR) is no longer the exception used only in a modest number of the largest cases, but instead starts to become the norm for complex litigation.

It is noteworthy that, whilst the use of predictive coding was contested in this case, it was proposed by the party giving the disclosure in question and opposed only by a party with no involvement in that process with little disclosure to give of his own.A bigger test will come when the position is reversed and a party seeks to impose the use of predictive coding or other form of TAR upon an unwilling disclosing party.Whilst it is difficult to envisage a situation in which a disclosing party would not wish to avail itself of the potential benefits of using some form of TAR in suitable cases, there could conceivably be concerns around highly sensitive documents and/or privilege which are perhaps more difficult to resolve without employing more rudimentary methods of review to a greater extent.There could also be tactical advantages to a disclosing party with deep pockets by seeking to make the costs generated by all sides during disclosure as high as possible.In any event, if in such circumstances the court follows the overwhelming endorsement of predictive coding given in Pyrrho and BCA Trading, it may well order a disclosing party to use predictive coding or other form of TAR against its wishes. That would be a further significant step towards making the use of TAR the norm rather than the exception.

Finally, also of significant interest is the court's willingness in this case to have the parties engage in a collaborative process of identifying and reducing issues before embarking on disclosure. It makes sense for parties to have a clear view of what is actually needed by way of disclosure before embarking on the process of searching for and reviewing documents, rather than the other way around. Significant time and costs savings could result from a greater focus on these matters at an early stage, so long as the discussions are conducted reasonably.