Running an E-Discovery Project: The Practical Workflow Behind the Technology

What e-discovery actually involves

E-discovery (or e-disclosure, the term more commonly used in England and Wales) is the process of identifying, collecting, reviewing and producing electronic documents that are relevant to litigation, investigations or regulatory inquiries. Email, messaging apps, shared drives, cloud storage and mobile devices all fall within scope. The volume involved in even a mid-sized commercial dispute can run into hundreds of thousands of documents, which is why the process has become a discipline in its own right rather than something fee earners do by hand.

The widely used reference model for the process is the Electronic Discovery Reference Model (EDRM), which breaks the work into stages: information governance, identification, preservation, collection, processing, review, analysis, production and presentation. Firms and their technology providers organise projects around this structure because it gives everyone involved, lawyers, clients and the court, a common language for what has been done and what remains outstanding.

Identification, preservation and collection

Before any review happens, the parties need to work out where relevant data might exist. This stage involves interviewing custodians (the individuals whose documents matter), mapping systems such as email servers, HR platforms and messaging tools, and issuing a legal hold notice instructing staff not to delete anything that might be relevant. Getting this wrong, by failing to preserve data or by collecting too narrowly, can lead to serious problems later if material turns out to have been lost.

Collection itself is usually done by specialist forensic teams using tools that capture data without altering metadata such as creation dates, authorship and file paths. Metadata matters because it can be as evidentially significant as the content of a document itself, for example in establishing when a file was created or who last edited it.

Processing and culling the data

Raw collections are enormous and full of duplicate, irrelevant or system-generated material. Processing software extracts text and metadata, converts files into a reviewable format, removes exact duplicates (deduplication) and can group near-identical documents together (email threading) so reviewers see a conversation once rather than fragment by fragment. Search terms, date ranges and custodian filters are then applied to cull the dataset down to something that human reviewers can realistically work through. This culling stage is where most of the volume reduction happens, often before anyone has opened a single document for substantive review.

Review: human, assisted and automated

Once the dataset is narrowed, documents are reviewed for relevance, privilege and confidentiality. Three broad approaches exist, and most real projects blend them:

• Linear review, where teams of reviewers go through documents in batches, remains the most straightforward but is slow and expensive at scale.
• Technology-assisted review (TAR), sometimes called predictive coding, uses machine learning trained on decisions made by senior lawyers on a sample set of documents. The software then ranks or categorises the remaining documents by likely relevance, allowing reviewers to focus effort where it matters most.
• Analytics tools such as concept clustering, email threading and communication pattern analysis help reviewers navigate large datasets without reading every document individually.

TAR does not remove lawyers from the process. It is calibrated and checked by experienced reviewers throughout, with statistical sampling used to validate that the model is performing to an acceptable standard before it is relied upon.

Cost, proportionality and project management

E-discovery is expensive, and cost is usually the main driver of how a project is scoped. Civil procedure in England and Wales requires disclosure to be proportionate to the value and complexity of the claim, so parties typically agree a disclosure review document setting out search terms, date ranges, custodians and the review methodology before work begins. This is as much a negotiation and project management exercise as a legal one: technology vendors, e-discovery consultants, in-house teams and external lawyers all need to coordinate on timelines, budgets and quality control.

Most firms now use dedicated e-discovery platforms, either run in-house or through specialist providers, rather than general document management systems. Choosing a platform involves considerations around data security, hosting location (particularly for data that must stay within certain jurisdictions), audit trails and the ability to defend the methodology if challenged later.

Production and presentation

Once review is complete, relevant and non-privileged documents are prepared for production, often with redactions applied to protect privileged or confidential content, and loaded into a format the receiving party or tribunal can use. In larger disputes, documents may later need to be presented at trial through electronic bundling systems, which is a separate but related skill set.

Why this matters to clients

For clients, understanding this workflow helps explain why disclosure costs can rival or exceed other litigation spend, and why early engagement with technology and a clear scope agreement tends to save money rather than cost it. Firms that invest in strong e-discovery processes are generally able to manage large disputes more efficiently and defend their methodology if it is ever questioned by the other side or the court.

Anyone wanting the procedural rules governing disclosure in England and Wales, including the practice direction dealing specifically with electronic documents, should consult the Civil Procedure Rules directly, since the detailed requirements are periodically updated.