Document Review
Document review is likely the most discussed phase of the EDRM model and the most discussed activity in the ediscovery life cycle. There are a few reasons why there’s so much attention focused on review.
It’s the phase that typically involves the most legal expertise to make responsiveness and privilege determinations (among other potential legal decisions made).
Document review has historically been the most expensive phase in ediscovery.
Typically, review is the primary activity for determining which documents will be produced, which ones will be withheld, and why.
Due to these factors, the methodology for how document review is conducted tends to be the issue that is most negotiated between parties in litigation, with debates ranging from defensibility to proportionality of the process.
This complexity underscores the need for a comprehensive understanding of various data formats and their implications for review. The increasing integration of artificial intelligence and machine learning tools also means that the review process is constantly being reshaped, offering both unprecedented efficiencies and new considerations for defensibility and transparency.
These advancements, coupled with the ever-present need to balance thoroughness with proportionality, make effective document review a dynamic and multifaceted discipline.
The Importance of Document Review
Document review is an essential phase of ediscovery, acting as the key link that connects collected data to legal strategy. It involves examining collected documents, typically ESI, to determine their relevance to the case, responsiveness to discovery requests, and whether they are privileged or confidential. In practical terms, document review is the human-driven process that distinguishes important information from extraneous data following collection and processing. This stage ensures that only the required, non-privileged information is disclosed to the opposing party, while sensitive or irrelevant material is filtered out in accordance with legal obligations.
Role in Legal Discovery
In litigation and investigations, document review is considered a fundamental part of discovery. It is the step where attorneys identify important pieces of evidence that will influence the legal proceedings. By examining a large amount of information to find relevant facts, lawyers can narrow the issues, find support for their claims or defenses, and eliminate irrelevant content. Thorough review prevents unexpected developments by ensuring both sides know the key evidence in advance.
Under the Federal Rules of Civil Procedure, parties must produce all non-privileged, relevant information requested by opponents (within proportional limits) – a duty that requires careful document review in order to comply with discovery rules and to ensure fair outcomes.
Impact on Compliance and Case Strategy
Document review isn’t just a procedural step in the process – it also directly impacts case strategy and compliance. For instance, review teams are required to identify and flag attorney-client privileged documents to ensure they are withheld and properly logged, thereby safeguarding client confidentiality. Failing to do so can result in waiver of privilege (for example, in one case, a party waived privilege on 613 documents due to their inadvertent disclosure).
Review is also important for regulatory and internal compliance matters: during regulatory investigations or audits, organizations examine data to find information relevant to government inquiries while meeting compliance requirements. The information obtained during review can inform legal strategy – attorneys may identify key emails that support their case or notice patterns that indicate a need to settle early.
Benefits of Efficient Document Review
Improving the efficiency of document review offers significant advantages. A faster, well-organized review process enables legal teams to meet tight deadlines (for productions or investigations) and reduce overall litigation costs and time. It also allows attorneys to concentrate on analysis and case building rather than handling manual tasks.
An efficient review process ensures consistency, enhancing quality and defensibility. It also reduces the risk of human error – technology and careful workflows can prevent important emails from being overlooked or confidential files from being accidentally produced.
The Cost of Document Review
Over the past 15 years, document review has consistently been identified as the most expensive and resource-intensive component of ediscovery. Several industry studies have demonstrated that reviewing documents consumes the majority of discovery expenditures.
For instance, a groundbreaking RAND study found that the review phase accounted for approximately 73% of the total costs of producing documents during ediscovery. In contrast, data collection represented roughly 8%, and processing constituted about 19% of costs in that analysis.
More than a decade later, review remains the largest cost driver. By 2024, according to ComplexDiscovery, an estimated 64% of ediscovery spending was still devoted to document review tasks, with an estimate that it will drop to 52% by 2029. While this is progress, it illustrates how review has consistently dominated the budget, despite changes in technology and workflows.
The biggest reason for that is that the volume of data subject to discovery has significantly increased over the past 15 years, initially leading to higher costs. According to Statista, data in the world has grown from 2 zettabytes (2 billion terabytes) in 2010 to 182 zettabytes by 2025 and is expected to continue to grow to 394 zettabytes by 2028!
The volume increase has been driven by a much greater variety of ESI sources, including emails, files, text and chat messages, and other ESI in each case resulted in more documents requiring review. Initially, organizations attempted to address this issue by increasing the number of human reviewers, which caused expenses to rise dramatically.
The expansive growth of data in organizations has forced legal teams to look for ways to leverage technology (such as predictive coding and generative AI) to keep up with data growth, as it has made cost management strategies in document review more important than ever for legal teams. However, those efforts have been countered by opposing counsel requesting more transparency into these technology leveraged review methods to ensure that the process is defensible. That’s why the approach to document review is one of the most common ediscovery disputes that parties encounter during discovery.
Potential Challenges and Issues Associated with Document Review
Document review in ediscovery presents numerous challenges, particularly as data volumes and types continue to expand. Critical issues include the substantial volume of data, the complexity and diversity of documents, the limitations of human reviewers (accuracy and error rates), the necessity to protect privileged and confidential information, regulatory hurdles, and ongoing cost and resource pressures.
Volume and Complexity of Documents
As discussed above, the amount of ESI potentially subject to discovery has significantly increased. Without effective culling and organization, the volume and complexity of modern data can challenge a review project.
Modern organizations manage a wide range of data, including emails, office documents, text messages, chat logs, social media posts, and collaboration platform data (e.g., Slack, Teams).
Cases today frequently involve tens of terabytes of data or millions of documents.
The sheer volume and variety of data make traditional linear review approaches impractical and cost-prohibitive for manual review.
Data exists in diverse formats, including emails, office files, text/chat messages, audio and video files, and even generative AI content, adding complexity to the review process.
Different data formats require specific processing steps before review, such as OCR for scanned images, email threading, and audio transcription.
Reviewers must navigate non-traditional communication elements, such as conversation threads, social media slang, emojis, and code words, which differ from standard business correspondence.
Sophisticated review strategies are necessary to manage data complexity, requiring tools to normalize and organize information efficiently.
Review teams need appropriate technology to handle the diversity of modern data formats.
Accuracy and Risks of Human Error
Document review has historically been a human process, accompanied by the risk of errors and inconsistencies. Human reviewers experience fatigue due to the monotony of examining hundreds of documents daily, and different individuals may apply different judgments to the same documents. Such variability implies that relying solely on human judgment can lead to the omission or misclassification of critical information. Errors such as failing to identify truly relevant documents (false negative) or marking too many irrelevant ones as relevant (false positives) can lead to legal disputes over missing key evidence or inundating the opposing party with a “document dump” of irrelevant data.
Inconsistencies can also complicate subsequent tasks; for example, if one reviewer designates a document as “Privileged” while another classifies a similar document as “Not Privileged,” there is a risk of unintentional disclosure. Additionally, as document volume increases, the likelihood of human error within the set rises. Even the most diligent teams will make errors.
Quality concerns have spurred interest in machine learning assistance, as algorithms, once trained, consistently apply criteria across the entire dataset without becoming fatigued or careless. Of course, the training is not always straightforward, and it often is conducted iteratively until a desired accuracy rate is achieved. Either way, humans remain integral to the process, necessitating the management of error through training, overlap (multiple reviewers verifying each other’s work), and sampling, to minimize accuracy concerns.
Managing Privilege and Confidentiality
Identifying privileged (attorney–client communications) or confidential information (trade secrets, personal data) is key to avoiding inappropriate disclosure. If privileged material is produced, doing so can waive protection and harm the client’s case. Mechanisms like clawback agreements and Federal Rule of Evidence 502(d) orders can mitigate inadvertent disclosures, especially if reasonable precautions were in place.
Flagging privileged documents consistently in large sets is challenging; reviewers need to recognize privileged threads and log them meticulously. Confidential business info or personally identifiable information (PII) may require redaction before production. Many teams create a privilege log, itemizing withheld documents and reasons, which requires significant effort. Extreme diligence is often needed to minimize the potential for privilege documents to slip through the cracks, so some firms use dedicated privilege review teams for thorough screening, but this can drive up costs. Reviewers must balance protecting sensitive info and speed of review, as overly aggressive privilege flagging can draw challenges from opposing counsel. The complexity of modern communications has made this process more nuanced and riskier than ever.
Regulatory and Compliance Challenges
In addition to litigation, document review is also a phase in regulatory investigations, compliance audits, and cross-border disputes. These factors complicate document review projects, necessitating careful planning and often specialist advice (e.g., privacy counsel) to manage legal risks effectively.
Different regulations dictate data handling processes, requiring organizations to navigate various compliance requirements.
Data privacy laws, such as the General Data Protection Regulation (GDPR), impose strict guidelines on reviewing personal data, even for litigation purposes.
Companies may need to anonymize data or obtain special permissions to transfer personal data across borders.
Industry-specific regulations, such as the Financial Industry Regulatory Authority (FINRA) for broker-dealers and the Health Insurance Portability and Accountability Act (HIPAA) for healthcare data, require tailored data handling approaches.
Review teams must consider these regulations, potentially segregating sensitive data or utilizing secure review facilities.
Regulatory investigations demand rapid and thorough responses, often requiring all relevant documents to be produced under tight deadlines.
The principle of proportionality, which limits review scope to reasonable case-specific extents, is often less flexible under government scrutiny, necessitating full compliance.
Maintaining audit trails and defensibility is crucial, as organizations must demonstrate that their review process was reasonable and legally compliant.
Documentation of reviewer training, quality control measures, and decision-making regarding responsiveness and privilege is essential for defensibility in court.
Cross-border ediscovery adds complexity, requiring organizations to navigate conflicting privacy and secrecy laws across multiple jurisdictions.
Companies involved in international litigation must balance discovery obligations with regional privacy laws, potentially filtering personal data or conducting reviews within the country of origin.
Costs and Resource Allocation Concerns
All of these challenges contribute to the broader issue: review processes are costly and resource-intensive. Organizations frequently contend with allocating adequate resources (personnel, time, money) to conduct comprehensive reviews without exceeding budgetary constraints. There is persistent pressure to achieve more with fewer resources and review increasing volumes of data more rapidly and economically. Failure to do so creates a tension between thoroughness and cost-effectiveness. For smaller cases or organizations with limited resources, there is a risk of not allocating enough resources to the review process, potentially leading to overlooked information or errors. Conversely, larger cases may have sufficient resources but face logistical challenges in coordinating extensive review teams.
Securing skilled reviewers presents another challenge, as document reviews are frequently conducted by junior attorneys or contract lawyers who may lack extensive experience and effectively supervising them poses its own difficulties. Balancing quality versus quantity is an ongoing concern – reviewers are typically under time constraints, yet hasty efforts can result in mistakes.
Review processes also necessitate robust coordination from a project management standpoint, with review managers assigning documents in batches, ensuring the team adheres to deadlines (including court-mandated production deadlines), and reallocating resources as needed to address unexpected developments, such as a late data influx or realizing mid-review that certain searches were insufficient.
Given these complexities, document review should be approached strategically, utilizing technology and processes to alleviate these challenges. Primary challenges and issues like the volume of evidence, complexity of the review process, human fallibility, sensitivity of information, and resource constraints must be addressed to ensure a well-managed document review project.
Finding the Right Technology
Selecting appropriate technology for document review is an important decision that can affect the efficiency, accuracy, and security of the ediscovery process. In recent years, the market for ediscovery review platforms has grown, and tools have advanced to include features like artificial intelligence, automation, and cloud-based collaboration. When assessing document review technology, legal teams should consider several key features and capabilities, compare popular platforms, and use a structured approach to determine which solution best fits their specific needs.
Predictive Coding and Generative AI
Predictive coding and GenAI are critical components of an advanced ediscovery document review solution, enabling legal teams to efficiently manage vast volumes of data while maintaining accuracy and compliance. Predictive coding leverages machine learning algorithms to identify and prioritize relevant documents based on patterns learned through training from human reviewers, significantly reducing the time and cost associated with manual review.
In addition to also classifying documents, GenAI enhances review by automating document summarization, extracting key insights, and generating contextual explanations that aid in legal analysis. Together, these technologies improve review speed, ensure consistency, and help legal teams focus on strategic decision-making rather than repetitive tasks.
Advanced Search and Analytics
The software should also include comprehensive search capabilities, including advanced querying (Boolean logic, proximity, wildcard, etc.), concept or fuzzy search to identify variations (such as misspellings), and filtering by metadata (dates, custodians, file types). Modern platforms may include analytics tools that group or visualize data automatically – for example, email threading (connecting emails with their replies and forwards), near-duplicate detection (identifying textually similar documents), and concept clustering (organizing documents by topic). These features enable a more strategic review of the data: entire email conversations can be reviewed together, or a representative from a set of near-duplicates can be handled instead of each copy.
Features like data visualization dashboards or communication network graphs are useful for exploring patterns in the data. Such analytics aid in early case assessment and help reviewers focus on the most relevant data. In summary, the software should provide multiple methods to organize the document set, enabling a more efficient and intelligent review process.
Integration and Data Handling
It’s also important to ensure the review platform integrates well with other ediscovery stages and data sources. It should ingest processed data and preserve all metadata and family relationships. Modern solutions often offer end-to-end capabilities or APIs/connectors to import data from enterprise sources like Office 365, Google Workspace, and Slack.
Integration capabilities are necessary for direct data collection from cloud repositories and exporting reviewed documents to other systems. The platform must also scale efficiently: it should handle large document volumes without performance issues, with auto-scaling for cloud-based solutions or efficient hardware utilization for on-premise setups.
Security and Compliance
Given the sensitive nature of reviewed documents, strong security measures are essential. The platform should offer encryption (at rest and in transit), user access controls, and audit trails. Look for providers with security certifications like ISO 27001 and SOC 2 Type II, as well as Transport Layer Security (TLS) version 1.2 for encryption in transit and AES-256 for encryption at rest.
For specific industries or government work, ensure compliance with standards like FedRAMP and StateRAMP. Features like PII detection and redaction tools enhance compliance. A secure platform protects against breaches and instills confidence during ediscovery processes. Carefully vet the provider’s security track record and capabilities.
User-Friendliness and Workflow Management
The tool's ease of use is vital for effective team performance. A user-friendly interface with a clean layout, logical menus, and simple tagging can boost productivity and cut training time. Look for features like document preview, bulk coding, and keyboard shortcuts. Consider if the platform offers customizable layouts, note-taking on documents, and workflow management, such as assigning document batches, tracking progress, and quality control checks.
Collaboration and Audit Features
Document review typically involves a team, so consider features that support collaboration among reviewers. This could include the ability for multiple people to review or comment on the same document, tagging workflows that notify another team member (e.g., a reviewer flags a document for second-level review), or integrated communication tools (such as chat or annotation systems for internal discussions). For remote or distributed teams, having everyone on a unified platform with real-time updates is beneficial.
Additionally, robust audit trail and reporting features assist with project management and defensibility. The software should log who accessed or coded each document and when, and be able to generate reports (e.g., lists of all documents tagged privileged, or statistics on reviewer agreement rates). These audit logs can be important if questions arise about actions taken (for example, to show that a document was marked as privileged and never produced, in case of a dispute). Effective reporting tools also help measure efficiency – tracking each reviewer’s throughput or identifying if certain search terms are yielding too many irrelevant hits.
Preparing a Document Review Strategy
An effective document review strategy is necessary for managing the review process efficiently and defensibly. Without a plan, you risk wasting time, getting inconsistent results, and making legal errors. Establishing clear procedures, leveraging best practices, and smartly using tools will optimize your review workflow. Below are best practices and considerations for planning and executing a successful document review strategy:
Early Planning and Scoping
It’s important to clearly define the review scope. The legal team should understand the key issues, claims, defenses, and what makes a document relevant. Identify date ranges, main custodians, and keywords or topics of interest. This will guide your search terms and prioritization, avoiding over-collecting unnecessary material. Set goals based on production deadlines and set milestones for review completion. Early planning may involve performing Early Case Assessment (ECA) by analyzing a sample of the data to refine your approach. Mapping out data contours and the needs of the case early helps allocate resources effectively and prevents surprises.
Developing a Review Protocol
Consistency is important in document review, so creating a written review manual is usually a good idea. This protocol should define how to categorize documents, including tags like “Responsive/Relevant”, “Non-responsive”, “Privileged”, and confidentiality levels. Additional tags such as “Key/Hot” or by issue can also be helpful.
Decide on privilege tagging and logging – some reviewers might apply a “Privileged” tag and provide necessary details for the privilege log. Include instructions on handling duplicates and emails with attachments. Establish rules for edge cases like partially privileged documents (e.g., redact or withhold entirely). A well-crafted protocol helps ensure consistency across all reviewers.
Workflow Optimization
It’s important to plan the workflow in phases or logical steps. That includes potentially starting with a first pass to identify obvious non-responsive documents, followed by a detailed second pass on closer calls. You might also organize the review by custodians or time periods, assigning teams accordingly.
If using TAR, your workflow could include developing a sampling seed set, training the model, conducting continuous review with guidance, and reviewing a validation sample at the end to ensure completeness. Even without TAR, that random validation sample can validate that non-relevant sets don’t contain hidden relevant documents. It’s also important to consider prioritizing certain document sets, such as key executive emails first while system files or peripheral documents may be reviewed later. By prioritizing, you can ensure at least the most important documents are reviewed by key deadlines.
Leveraging Technology and Analytics
Consider integrating technology early, including the use of predictive coding for large cases to prioritize documents, starting with a quick training phase where senior reviewers label samples and the algorithm ranks documents by relevance. Focus on important documents first for early insights and potentially reduce review time. Even without TAR, use analytics like email threading and deduplication – enable these features on review platforms to streamline the process by reviewing inclusive emails only.
Leveraging clustering or concept analysis can reveal thematic groups in your data, highlighting unexpected topics. Test and refine search terms to filter out non-relevant documents. Consider combining search term filtering with TAR for more efficient document review. Use analytics for quality control where appropriate. It’s important to plan up front how to incorporate these tools seamlessly into your workflow.
Team Coordination and Training
Effective strategy relies on both tools and people. Assemble a skilled team with defined roles: senior attorneys or experts for core tasks, junior or contract reviewers for initial reviews, and (if necessary) separate members for privilege review. Provide necessary training on review protocol and case background to ensure they understand the case specifics, important facts, and what to look out for. Familiarize them with the review platform and clarify any uncertainties through Q&A channels. Promote collaboration to maintain consistency and efficiency, ensuring significant findings are shared. Establish a review schedule and targets to keep the project on track and accurately forecast completion.
Quality Control and Defensibility
Your strategy should include implementing quality control (QC) measures throughout the review. A second-level review can help catch errors, where a percentage of each reviewer’s documents are checked by a senior reviewer or a different team member. Significant disagreements or errors should lead to corrections and retraining if needed. Track agreement rates and metrics, such as how often second-level reviews change first-level decisions, to identify issues.
Use searches to double-check for key terms that should have been tagged or privileged content among non-privileged documents. Utilize random sampling of non-responsive documents to ensure nothing important was missed. Ensure all privilege-tagged documents undergo QC by a senior attorney. Document all QC steps to demonstrate defensibility. A strong QC process minimizes errors and provides confidence in the accuracy and completeness of the production.
Communication and Case Integration
Maintain open communication between the review team and the broader case team. Establish mechanisms like a "hot doc" folder or regular update meetings to highlight significant documents for attorney review. This enables real-time adjustments to case strategy based on new evidence. For instance, discovering a "smoking gun" email can prompt changes in deposition plans or settlement posture.
Attorneys may also refine their focus as new information emerges. Regular touchpoints with the legal team help keep review objectives aligned with the evolving case strategy and motivate the review team by highlighting the importance of their findings.
Balancing Speed and Thoroughness
Time constraints are common in document reviews, but hurrying can cause errors. An effective strategy should balance these aspects. Techniques such as batching (assigning manageable portions to each reviewer) and setting daily/weekly targets can help the team progress efficiently. Tracking productivity is useful – which review platforms can help with by showing how many documents each reviewer codes per hour. If someone falls significantly below the expected rate, investigate if there is an issue (such as receiving large documents or needing guidance). Accuracy and completeness are critical.
Additionally, establish ground rules, such as: “If you are uncertain about a document, do not guess – flag it for discussion.” Pausing to resolve confusion often saves time compared to half the team guessing incorrectly and needing to redo work. Additionally, plan for straggler tasks – activities like privilege log assembly, redactions, and production formatting require time at the end. Ensure these are included in the timeline to prevent them from being rushed (for example, don’t schedule the review to end on the production due date; finish the review earlier to allow time for final tasks).
Document Review Checklist
When starting a document review project, it’s helpful to follow a step-by-step plan to ensure all bases are covered. While each review project is different, there are certain steps you should complete when applicable to your specific review. Below is an example of a checklist of key steps and milestones to guide an ediscovery document review from start to finish, incorporating efficiency measures and compliance checks at each stage:
1. Define the Scope and Objectives
Clearly define the purpose of the review and what it needs to accomplish.
Identify the universe of data to be reviewed (e.g., number of documents, gigabytes, sources, custodians).
Determine relevant date ranges and subject matter.
Establish legal issues or requests the documents must respond to.
Set clear timelines, including production deadlines and interim milestones.
2. Assemble the Team and Resources
Decide if the review will be in-house or outsourced.
Select a legal service provider if outsourcing (ensure NDAs are in place).
Assign team roles, including first-level reviewers and senior attorneys for QC.
Ensure access to subject matter experts as needed.
Select and prepare the review platform (licenses, data loading, analytics setup).
Allocate IT support for troubleshooting software or data issues.
3. Establish Review Protocols and Guidelines
Define relevance and responsiveness criteria.
Outline document tagging categories (e.g., Responsive, Privileged, Confidential).
Specify rules for handling families (emails and attachments), duplicates, and threads.
Set up a privilege log template if required.
Document and circulate review guidelines to all team members.
4. Prepare the Data for Review (Culling and Processing)
Ensure all collected data is processed and de-duplicated.
Apply filters to eliminate clearly irrelevant data (e.g., date ranges, custodians, file types).
Use search terms to exclude non-responsive documents while ensuring validation.
Document all culling decisions for defensibility.
Confirm data is loaded in the review tool and ready for review.
5. Conduct Team Training and Kickoff
Train reviewers on the review protocol and provide case background.
Explain key players, allegations, and important document types.
Provide platform training on document viewing, tagging, and searching.
Clarify how to interact with AI-assisted tools, if applicable.
Establish workflow procedures, including batch allocation and issue escalation.
Set up communication channels for questions and clarifications.
6. Execute the Document Review Workflow
Assign batches to reviewers and track progress daily.
Monitor for bottlenecks or technical issues.
Maintain ongoing communication and address common reviewer questions.
Optimize workflows using AI tools (e.g., TAR) if applicable.
Periodically assess progress (e.g., 25%, 50%, 75% completion milestones).
Adjust resources or scope if review falls behind schedule.
7. Quality Control and Second-Level Review
Perform interim QC checks rather than waiting until review completion.
Have senior reviewers check a sample of reviewed documents for consistency.
Run searches on non-responsive documents to catch potential mis-tagging.
Verify privilege calls and ensure accuracy of the privilege log.
Update review protocol based on QC findings and retrain reviewers if needed.
Conduct final QC sampling to confirm accuracy before production.
8. Production and Closure
Extract and prepare the final production set (responsive and non-privileged documents).
Apply redactions where necessary.
Verify the completeness and accuracy of the production set.
Generate and finalize the privilege log.
Ensure production format meets required specifications.
Document and confirm the final number of documents produced and withheld.
Secure and archive data as per retention policies.
9. Post-Review Debrief and Lessons Learned
Conduct a post-review meeting to discuss successes and challenges.
Evaluate effectiveness of filtering, AI tools, and reviewer accuracy.
Capture lessons learned for future improvements.
Update internal review playbooks with refinements made during the review.
Ensure review insights are handed off to the legal team for trial or investigation.
Confirm all obligations related to the review have been met (e.g., regulatory compliance).
By following this checklist, you can systematically approach a document review project and minimize the risk that something important is overlooked. Each step is key to achieving a defensible, efficient review that meets legal requirements and deadlines.