Case Details
- Citation: [2013] SGHC 111
- Court: High Court of the Republic of Singapore
- Decision Date: 22 May 2013
- Coram: Lee Seiu Kin J
- Case Number: Suits Nos 913 and 914 of 2009 (Registrar's Appeals Nos 421 and 422 of 2012)
- Hearing Date(s): 24 April 2013
- Claimant / Plaintiff: Global Yellow Pages Limited
- Respondents / Defendants: Promedia Directories Pte Ltd (Suit 913); Streetdirectory Pte Ltd (Suit 914)
- Counsel for Claimant: Karen Teo, Adeline Chung and Han Hsien Fei (TSMP Law Corporation)
- Counsel for Respondent (Suit 913): G Radakrishnan (Infinitus Law Corporation)
- Counsel for Respondent (Suit 914): Zhulkarnain Abdul Rahim and Diyanah Binte Baharudin (Rodyk & Davidson LLP)
- Practice Areas: Civil Procedure; Discovery of documents; E-discovery; Copyright Infringement
Summary
The judgment in Global Yellow Pages Limited v Promedia Directories Pte Ltd and another suit [2013] SGHC 111 represents a seminal exploration of the procedural mechanics governing electronic discovery (e-discovery) within the Singapore legal landscape. The dispute arose from two consolidated copyright infringement actions initiated by Global Yellow Pages Limited (“GYP”) against its competitors, Promedia Directories Pte Ltd (“Promedia”) and Streetdirectory Pte Ltd (“Streetdirectory”). While the underlying merits concerned the alleged unauthorized reproduction of directory data and the use of fictitious “Seeds” to detect copying, the specific focus of these Registrar’s Appeals was the Assistant Registrar’s (“AR”) order regarding the implementation of keyword searches on electronic devices.
The High Court was tasked with resolving the tension between the traditional “leave no stone unturned” approach to discovery and the modern necessity for proportionality in the digital age. Lee Seiu Kin J affirmed that the discovery of electronic documents requires a departure from manual inspection methods, which are “enormously impractical” given the sheer volume of data stored on modern devices. The court endorsed a conceptual framework where keyword searches are viewed not as a one-off event, but as an iterative sieving process. This process requires parties to collaborate and refine search terms to balance the need for comprehensive disclosure against the risk of generating excessive “noise” or irrelevant hits.
In dismissing the appeals with minor variations, the court emphasized that the selection of keywords must be grounded in the pleaded issues. The inclusion of the term “seed” or “seeds” was upheld as relevant because it directly related to GYP’s factual theory of the case—namely, that the defendants had copied fictitious listings. However, the court also demonstrated a willingness to intervene where protocols were overbroad or vague, imposing a specific time limit on the “seed” search and striking out a keyword that functioned merely as a placeholder for custodian names. This decision serves as a primary authority for practitioners on how to structure e-discovery protocols that satisfy the requirements of Order 24 Rule 1 of the Rules of Court while maintaining cost-effectiveness.
Ultimately, the judgment reinforces the principle that “perfect justice” in discovery is an elusive and potentially self-defeating goal if it leads to prohibitive costs. By adopting the reasoning in Nichia Corporation v Argos Ltd [2007] EWCA Civ 741, the Singapore High Court signaled a clear preference for a pragmatic, technology-assisted approach to discovery that prioritizes the fair and efficient disposal of litigation over exhaustive, manual document review.
Timeline of Events
- 2009: Global Yellow Pages Limited commences Suit No 913 of 2009 against Promedia Directories Pte Ltd and Suit No 914 of 2009 against Streetdirectory Pte Ltd, alleging copyright infringement in its directories.
- 17 September 2010: Freddie Tan Poh Chye files an affidavit in Suit 914 (“Freddie Tan’s Affidavit”), which refers to the “Seeds” as defined in GYP’s pleadings.
- 13 December 2010: Parties reach an impasse on the discovery process, leading to further procedural interventions (as noted in the context of related e-discovery jurisprudence).
- 12 January 2011: Continued procedural management of the discovery applications and the refinement of electronic discovery protocols.
- 17 February 2012: The Assistant Registrar hears the discovery applications (SUM 774/2011 and SUM 773/2011) and issues orders for the defendants to conduct keyword searches on specified electronic devices.
- 27 March 2012: Promedia and Streetdirectory file Registrar’s Appeals (RA 421/2012 and RA 422/2012) against the AR’s discovery orders.
- 6 November 2012: A significant date identified by the court as the end-point for the restricted “seed” keyword search, corresponding with the period after which the term “seed” became a known concept in the litigation.
- 24 April 2013: The High Court hears the arguments for RA 421/2012 and RA 422/2012.
- 22 May 2013: Lee Seiu Kin J delivers the judgment, dismissing the appeals with two minor variations to the AR’s orders.
What Were the Facts of This Case?
The plaintiff, Global Yellow Pages Limited (“GYP”), is a Singapore-incorporated company listed on the Singapore Exchange. GYP’s primary business involves the publication of directories and the provision of classified directory advertising. The defendants are competitors in the directory and location-based services industry. Promedia Directories Pte Ltd (“Promedia”) publishes directories, including “The Green Book,” while Streetdirectory Pte Ltd (“Streetdirectory”) develops location-based software and solutions, including an online directory.
GYP alleged that it owned the copyright in various directory compilations, including printed editions and an online “Internet Yellow Pages” directory. GYP claimed that these directories were original intellectual creations resulting from the selection and arrangement of subscriber information, which GYP had verified and classified. The core of GYP’s complaint was that Promedia and Streetdirectory had infringed this copyright by reproducing substantial portions of GYP’s listings in their own products—specifically Promedia’s “The Green Book” (in print and CD-ROM) and Streetdirectory’s website.
To prove copying, GYP relied on the presence of “Seeds”—fictitious subscriber listings that GYP had intentionally planted in its directories. GYP argued that because these listings were fake, their appearance in the defendants’ directories could only be explained by direct copying. The defendants denied infringement, challenging the originality of GYP’s compilations and asserting that they had independently created their own databases. Promedia specifically pleaded a defense of fair dealing under s 35 of the Copyright Act (Cap 63, 2006 Rev Ed), as well as defenses of public interest, laches, and acquiescence.
The procedural dispute centered on GYP’s application for discovery of electronic documents from the defendants. GYP sought to uncover internal communications, drafts, and databases that might reveal the source of the defendants’ listings. Given the digital nature of the defendants’ businesses, the discovery necessarily involved searching computers, servers, and other electronic storage media. The parties had attempted to agree on an electronic discovery protocol, but remained in dispute over the specific keywords to be used and the scope of the search.
The Assistant Registrar had ordered that the defendants conduct searches using a list of keywords. These keywords included terms like “seed”, “seeds”, “copy”, “follow”, and various deletion-related terms. The defendants appealed this order, arguing that the keyword searches were overbroad, irrelevant, and would result in the disclosure of a vast number of documents that had nothing to do with the alleged infringement. Streetdirectory particularly challenged the keyword “seed”, arguing it was a term of art used by GYP but not necessarily by the defendants in their internal processes. The defendants also objected to a keyword entry that required searching for the names of every custodian or employee involved in the creation of the directories, arguing it was not a proper “keyword” but a vague category.
The evidence record included an affidavit from Freddie Tan Poh Chye dated 17 September 2010, which became a focal point for the court in determining the temporal relevance of certain search terms. The court had to decide whether the AR’s selection of keywords was a proportionate response to the discovery needs of the case, or whether it imposed an undue burden on the defendants by requiring them to sift through thousands of irrelevant electronic files.
What Were the Key Legal Issues?
The appeals raised fundamental questions regarding the court’s supervision of the e-discovery process. The primary legal issues were:
- The Appropriateness of Keyword Searches: Whether keyword-based searching is a valid and necessary substitute for manual document review in cases involving large volumes of electronic data.
- The Definition of “Relevance” in E-Discovery: How the court should determine if a specific keyword is sufficiently connected to the pleaded issues to justify its inclusion in a discovery protocol.
- Proportionality and the “Iterative Process”: Whether the court should mandate an iterative approach to keyword selection—whereby preliminary searches are conducted to assess the number of “hits”—to ensure that the discovery burden is proportionate to the value of the information sought.
- Temporal Limitations on Search Terms: Whether the court should restrict the time period for specific keyword searches to prevent the retrieval of documents created before or after the term became relevant to the dispute.
- The Determinacy of Keywords: Whether a search term that acts as a placeholder (e.g., “[name of each custodian]”) is a valid keyword for the purposes of an e-discovery order.
These issues required the court to interpret the requirements of Order 24 Rule 1 of the Rules of Court in the context of modern information technology, balancing the plaintiff’s right to discovery against the defendants’ right to be protected from oppressive or unnecessary search requirements.
How Did the Court Analyse the Issues?
Lee Seiu Kin J began his analysis by acknowledging the “perennial tension” in civil procedure between the desire for exhaustive discovery and the need for proportionality. He cited his own previous observation in Breezeway Overseas Ltd and another v UBS AG and others [2012] 4 SLR 1035 at [20], noting that while the goal is to find the truth, the process must not become so expensive as to deny justice. The judge explicitly relied on the caution provided by Jacob LJ in Nichia Corporation v Argos Ltd [2007] EWCA Civ 741 at [50], where it was warned that a “leave no stone unturned” approach in the search for “perfect justice” can actually defeat justice by making litigation unaffordable.
The court then addressed the specific challenges of electronic discovery. Lee Seiu Kin J noted that the traditional model of discovery—where a solicitor manually inspects every document in a client’s possession—is “enormously impractical” for electronic data. He observed that the volume of data stored on a single hard drive can exceed the capacity of a traditional physical filing room. Consequently, the court accepted that keyword searches are a necessary tool to identify potentially relevant documents. However, the judge identified a critical problem: keywords can be either too broad (producing too many irrelevant “hits”) or too narrow (missing relevant documents).
To solve this, the court proposed the “Iterative Process”. Lee Seiu Kin J stated at [44]:
“In my view, the key to unlocking the full potential of keyword searches (bearing in mind the difficulties outlined above) is by way of adopting an iterative process.”
This process involves several steps:
- Parties propose an initial list of keywords based on the pleadings.
- Preliminary searches are conducted to determine the number of “hits” for each keyword.
- If a keyword produces an unmanageable number of hits, the parties must collaborate to refine the term (e.g., by using Boolean operators like “AND” or “NOT”, or by limiting the search to specific file types or time periods).
- The court intervenes only when parties cannot agree on the refinement.
Applying this to the keyword “seed” or “seeds”, the court found that these terms were highly relevant because GYP’s case was built on the allegation that the defendants had copied fictitious “seeds”. If the defendants had indeed copied these listings, it was likely that internal discussions about them would use that specific terminology. However, the court recognized the risk of overbreadth. The defendants argued that “seed” is a common word that could appear in many unrelated contexts. To balance this, the court looked at the timeline. The term “seed” became a central part of the litigation vocabulary around the time of Freddie Tan’s Affidavit in September 2010. Therefore, the court limited the search for “seed” to the period from 17 September 2010 to 6 November 2012. This temporal restriction ensured that the search targeted the period when the term was most likely to be used in a relevant context, while allowing GYP to apply for an expansion if they could later prove the term was used earlier.
Regarding the keyword “[name of each custodian/employee materially involved in creating, authoring and/or compiling the directories]”, the court took a stricter view. Lee Seiu Kin J found that this was not a “keyword” in the technical sense but a description of a category of people. In e-discovery, a keyword must be a specific string of characters that a computer can search for. A placeholder that requires a subjective determination of who was “materially involved” before the search can even begin is indeterminate and unworkable as a court-ordered keyword. Consequently, the judge ordered its deletion from the search protocol.
The court also considered the treatment of other keywords like “copy”, “follow”, and “delete”. While these are common words, they were deemed relevant to the allegations of unauthorized reproduction and the subsequent attempts to cover up such copying. The court maintained these keywords but emphasized that if they produced too many hits, the iterative process of refinement should be applied by the parties.
Finally, the judge addressed the role of the court in reviewing the AR’s discretion. He noted that while discovery orders are discretionary, the High Court will intervene if the order is based on a wrong principle or is clearly wrong. In this case, the AR had correctly identified the need for keyword searches but had not sufficiently narrowed the “seed” search or identified the flaw in the “custodian name” keyword. By making these “minor variations”, the High Court refined the AR’s order to better align with the principles of proportionality and determinacy.
What Was the Outcome?
Lee Seiu Kin J dismissed both Registrar’s Appeals (RA 421 and RA 422), largely upholding the Assistant Registrar’s discovery orders, but with two specific modifications intended to improve the proportionality and clarity of the e-discovery process. The operative order was stated at [3]:
“I dismissed both appeals except that I made two minor variations to the AR’s orders”
The two variations were as follows:
- Temporal Limitation on “Seed” Search: The search for the keywords “seed” or “seeds” on the electronic devices of the relevant custodians was restricted to the period between 17 September 2010 and 6 November 2012. The court granted GYP “liberty to apply” to expand this period if they could provide evidence that the term “seed” had entered the defendants’ vocabulary in a relevant context at an earlier date.
- Deletion of Indeterminate Keyword: The court deleted the entry “[name of each custodian/employee materially involved in creating, authoring and/or compiling the directories]” from the list of keywords. The judge determined that this was not a valid search term but a vague descriptor that failed the test of determinacy required for an automated electronic search.
Regarding costs, the court adopted a neutral stance given that the appeals were largely dismissed but resulted in some refinement of the orders. The judge ordered that the costs of the appeals be “in the cause”, meaning the ultimate winner of the main suits would recover these costs. However, he fixed the quantum of these costs to provide certainty for the parties:
“I also ordered that costs be in the cause and fixed the quantum of costs and disbursements at $1,000 in each suit.” (at [23])
The practical result was that the defendants were required to proceed with the e-discovery process using the refined keyword list and protocols, ensuring that the litigation could move forward toward a trial on the merits of the copyright infringement claims.
Why Does This Case Matter?
This case is a cornerstone of Singapore’s e-discovery jurisprudence. Its significance lies in the transition from traditional discovery principles to a technology-aware framework that prioritizes proportionality. For years, the “relevance” test in discovery was applied with a broad brush, often leading to “fishing expeditions” or oppressive disclosure requirements. Global Yellow Pages provides the necessary guardrails for this process in the digital era.
First, the endorsement of the “iterative process” is a major doctrinal contribution. It moves discovery away from being a combative, one-off exchange of lists and toward a collaborative, technical exercise. By requiring parties to conduct preliminary searches and report “hit” counts, the court forces a reality check on the scope of discovery. This prevents the “perfect justice” trap where the cost of finding a “smoking gun” exceeds the value of the claim itself. Practitioners must now approach e-discovery as a series of refinements rather than a single demand for all potentially relevant data.
Second, the case clarifies the standard for keywords. The court’s refusal to allow a placeholder for custodian names as a “keyword” is a vital lesson in technical precision. It signals that discovery orders must be capable of literal, automated execution. Lawyers cannot use discovery orders to outsource the identification of relevant personnel to a computer search; they must first identify the custodians and then search for specific, determinate terms within those custodians' data.
Third, the judgment reinforces the importance of the pleadings in scoping e-discovery. The court’s analysis of the term “seed” shows that a keyword’s validity is inextricably linked to the factual theory of the case. If a term is not part of the pleaded narrative, it is unlikely to survive a proportionality challenge. Conversely, if a term is central to the dispute, the court will allow it even if it is a common word, provided that temporal or Boolean filters are applied to minimize noise.
Finally, the case places Singapore firmly in line with international best practices, such as those seen in the UK (citing Nichia) and the Electronic Discovery Reference Model (EDRM). It demonstrates that the Singapore High Court is prepared to be hands-on in managing the technical aspects of litigation to ensure that the “saving of costs” limb of the overriding objective is met. For practitioners, this means that an e-discovery protocol is no longer a mere formality but a critical piece of litigation strategy that requires both legal and technical expertise.
Practice Pointers
- Conduct Preliminary “Hit” Searches: Before finalizing a discovery application or attending a hearing, counsel should conduct preliminary searches using proposed keywords to identify the number of hits. This data is essential for arguing proportionality.
- Use Boolean Operators to Refine: If a relevant keyword (like “copy”) is too broad, use “AND”, “OR”, or “NOT” connectors to narrow the results to the specific context of the dispute.
- Propose Temporal Filters: Always consider whether a keyword search should be limited to a specific date range. As seen with the “seed” keyword, limiting a search to the period when a term became relevant can significantly reduce the discovery burden.
- Ensure Keyword Determinacy: Keywords must be specific strings of text. Avoid using descriptions, categories, or placeholders (like “[names of employees]”) in a keyword list. Identify the individuals first, then list their names as specific search terms if necessary.
- Collaborate Early: The “iterative process” works best when parties cooperate. Attempt to agree on a search protocol that includes a mechanism for refining terms after the initial hit counts are known.
- Tie Keywords to Pleadings: Be prepared to justify every keyword by pointing to a specific paragraph in the Statement of Claim or Defence. If a keyword cannot be linked to a pleaded issue, it is vulnerable to being struck out.
- Document the Impasse: If the other side is being unreasonable in refusing to refine broad keywords, document the “hit” counts and the proposed refinements to demonstrate to the court that you have attempted the iterative process in good faith.
Subsequent Treatment
The ratio in Global Yellow Pages regarding the iterative process in e-discovery has been consistently followed in subsequent Singapore High Court decisions involving complex discovery. It is frequently cited alongside [2012] SGHC 41 to emphasize that discovery is a matter of proportionality and that courts will not hesitate to prune keyword lists that are likely to produce an oppressive amount of irrelevant data. The case remains the leading authority on the use of temporal limitations and the requirement for determinacy in search terms.
Legislation Referenced
- Copyright Act (Cap 63, 2006 Rev Ed): Specifically s 35 regarding the defense of fair dealing.
- Rules of Court: Order 24 Rule 1 (General Discovery) and the associated Electronic Discovery Practice Directions.
Cases Cited
- Considered:
- [2009] SGHC 194 (Fermin Aldabe v Standard Chartered Bank) - regarding the general process of discovery.
- Nichia Corporation v Argos Ltd [2007] EWCA Civ 741 - regarding the caution against “perfect justice” in discovery.
- Referred to:
- [2011] SGHC 61 (Robin Duane Littau v Astrata (Asia Pacific) Pte Ltd) - regarding impasses in discovery negotiations.
- [2012] SGHC 41 (Breezeway Overseas Ltd v UBS AG) - regarding the recommended iterative approach to e-discovery.
- Breezeway Overseas Ltd and another v UBS AG and others [2012] 4 SLR 1035 - regarding the perennial tension in discovery law.
- Sanae Achar v Sci-Gen Ltd [2011] 3 SLR 967 - regarding the problems of storing soft copies indefinitely.
- Digicel (St. Lucia) Ltd & Ors v Cable & Wireless Plc & Ors [2008] EWHC 2522 (Ch) - regarding pitfalls in failing to agree on search parameters.
Source Documents
- Original judgment PDF: Download (PDF, hosted on Legal Wires CDN)
- Official eLitigation record: View on elitigation.sg