Triumph Controls UK Ltd & Anor v Primus International Holding Co & Ors

Factual and Procedural Background

The claimants initiated proceedings seeking approximately US$65 million in damages from the defendants for alleged breaches of warranty following the sale of the defendants' aerospace business to the claimants via a share capital purchase. The claimants alleged that the defendants failed to disclose various operational failings and lack of accreditation at their facility located at Farnborough. The original trial date was adjourned due to the claimants' late application to re-amend the Particulars of Claim, with the trial rescheduled to commence on 18 June 2018 for five weeks.

In December 2017, the defendants applied for wide-ranging court orders addressing fundamental deficiencies they identified in the claimants' disclosure, including a request for the claimants to redo their document search exercise. Subsequently, the defendants narrowed their requests to two focused orders: (1) that the claimants provide the list of 860,000 folders and file paths identified on the Farnborough shared drive to allow the defendants to check for any missing folders or paths, and (2) that the claimants undertake a manual review of the remaining 220,000 documents (out of 450,000 identified by keyword search) which had not been fully searched beyond limited sampling.

The court held a hearing close to the trial date and issued oral conclusions, later promising a detailed judgment explaining the reasons. The judgment addresses the background to the disclosure dispute, applicable legal principles, and the two applications regarding the file paths and the 220,000 documents.

Legal Issues Presented

1. Whether the claimants' approach to disclosure of the Farnborough shared drive folders and file paths was reasonable and proportionate, and if further disclosure should be ordered.
2. Whether the claimants’ decision not to manually review the remaining 220,000 documents, relying instead on a sampling exercise using Computer Assisted Review (CAR), was adequate and proportionate, and if further manual review should be ordered.

Arguments of the Parties

Defendants' Arguments

• The claimants' disclosure was fundamentally deficient and required redoing the entire search exercise.
• The defendants requested access to the full list of folders and file paths to identify any omissions in the search.
• The defendants argued that the 220,000 documents, which had not been fully searched, should undergo manual review rather than relying solely on a CAR sampling exercise.

Claimants' Arguments

• The claimants initially conducted a keyword search agreed upon with the defendants, reducing documents to 450,000 responsive items.
• They manually reviewed approximately 230,000 documents aided by CAR, disclosing around 16,000 documents.
• For the remaining 220,000 documents, the claimants conducted a 1% CAR sample which predicted only 0.38% relevance, concluding further manual review was disproportionate and unnecessary.
• The claimants argued that their method of reducing folders and file paths was reasonable and proportionate and had been disclosed in their original List of Documents.
• They expressed concerns about the cost and time required for a manual review of the remaining documents.

Table of Precedents Cited

Precedent	Rule or Principle Cited For	Application by the Court
Nichia Corporation v Argos Limited [2007] EWCA Civ 741	Introduction of proportionality into disclosure rules; standard disclosure does not guarantee "perfect justice" but balances cost and justice pragmatically.	The court relied on this precedent to emphasise the balance between thoroughness and proportionality in disclosure, accepting that not every document can be disclosed if disproportionate.
Digicel (St Lucia) Limited and Others v Cable & Wireless Plc and Others [2008] EWHC 2522 (Ch)	Disclosure principles relating to electronic documents; unilateral decisions on keyword searches risk court ordering further searches; proportionality considerations in ordering further searches.	The court applied Digicel to critique the claimants' unilateral use of CAR and sampling without defendant input, highlighting the risk of inadequate disclosure and reinforcing the need for transparency and cooperation.
Pyrrho Investments Limited and Another v MWB Property Limited and Others [2016] EWHC 256 (Ch)	Guidelines on the use of Computer Assisted Review (CAR); importance of consistent application of relevance criteria and oversight by a senior lawyer.	The court referenced Pyrrho to underscore deficiencies in the claimants' CAR process, noting the absence of a senior lawyer overseeing the review and the lack of transparency in the sampling methodology.

Court's Reasoning and Analysis

The court began by examining the claimants' approach to the Farnborough shared drive, which contained approximately 20 million documents in 860,000 folders or file paths. The claimants reduced this to about one terabyte of data by having custodians identify relevant folders/file paths and cross-checking the full list. Although this was done without consulting the defendants, the court found that the claimants had been transparent about this process in their original disclosure documents. The defendants raised their complaint 17 months after becoming aware of the method, which the court deemed too late. The court found the claimants' approach reasonable and proportionate, noting that only two additional folders/file paths had been disclosed subsequently, a negligible amount. The defendants failed to identify any obviously missing folders or file paths. Accordingly, the court concluded that ordering further disclosure of folders/file paths was neither reasonable nor proportionate, dismissing this aspect of the defendants' application.

Regarding the 220,000 documents that had not been manually reviewed, the court expressed significant concerns. The claimants had promised a manual review of all documents responsive to the keyword searches in their Electronic Documents Questionnaire but did not fulfill this promise. Instead, they conducted a limited CAR-assisted review on approximately half the documents and a 1% CAR sampling on the remainder, predicting a low relevance rate of 0.38%. The court found this approach lacked transparency and independent verifiability, noting the absence of detailed information about the CAR setup, operation, and sampling methodology. The unilateral nature of the claimants' CAR use without defendant input exacerbated these concerns.

The court further criticized the CAR process for involving numerous paralegals and associates without apparent oversight by a senior lawyer, contrary to best practices outlined in Pyrrho. The court also highlighted the relatively low total disclosed documents (around 19,500) compared to the 450,000 responsive documents, suggesting potential under-disclosure. The fact that approximately 2,000 documents disclosed later should have been disclosed originally raised further doubts about the adequacy of the initial disclosure process. The court also noted that some disclosed documents were identified by witnesses and experts, indicating disclosure deficiencies.

Balancing these factors, the court concluded that the claimants' steps regarding the 220,000 documents were not adequate. Given the lack of transparency and the importance of proportionality, and with trial approaching, the court ordered the parties to agree on a methodology for a manual review of a 25% sample of these documents. This review was to be completed within three weeks, with results to be reported in an agreed letter to be presented at a subsequent hearing.

Holding and Implications

The court DENIED the defendants' application for further disclosure of additional folders and file paths on the Farnborough shared drive, finding the claimants' approach reasonable, proportionate, and sufficiently transparent given the circumstances and timing.

However, the court ORDERED that the claimants conduct a manual review of a 25% sample of the 220,000 documents that had not been fully searched, due to concerns about the adequacy, transparency, and unilateral nature of the prior CAR-assisted sampling. The manual review was to be completed within three weeks, with a report to be submitted to the court.

This decision directly affects the parties by imposing a limited additional disclosure obligation on the claimants, aimed at ensuring fairness and proportionality in disclosure ahead of trial. The court did not establish new precedent but applied existing principles emphasizing cooperation, transparency, and proportionality in electronic disclosure.