Appearances
Elliott Jung
Vol. 1September 11, 2026 – October 23, 2026Est. Featured
A New Era in Legal Journalism and Expertise
A New Era in Legal Journalism and Expertise
Q&A
Discovery in litigation often spirals out of control, consuming resources while missing critical evidence. This article explains how to maintain proportional scope without sacrificing case quality, featuring insights from legal professionals who have mastered the balance. Learn practical strategies for building case theory first to guide targeted discovery that actually moves cases forward.
Elliott Jung
Early in my career, I believed that the safest discovery request would be the most general request. If there were any possibility that a particular document would be relevant, I wanted to include it. However, real-life litigation made me change my view. Massive productions typically resulted in weeks of review time and obscured the facts that really determined the outcome of the dispute.
For cases involving catastrophic injuries or employment, I begin by analyzing the facts I need to prove my case, and build backwards from there. This requires analyzing the critical events that took place, the decision-makers involved, the communications in relation to those events, and an appropriate timeframe. Developing a theory of the case inherently leads to proportionate discovery. It also greatly simplifies explaining to a judge the reasonableness of a discovery request.
There is one particular approach that has saved clients much of their time and money. Identify the custodians and dates that matter most and then expand from there. If the documents lead somewhere, you can always request more. Typically, that approach is much faster than sifting through countless numbers of irrelevant documents.
After conducting more than 50 jury trials, I know that a jury cares about evidence, not the quantity of the production of documents. If you cannot justify requesting them, chances are you are asking for too much.
Tight search terms and smart deduping can cut cost without losing key facts. Test terms on a small set first and share hit counts and false positive rates to guide changes. Prefer proximity and date limits over broad words that pull noise.
Apply cross-custodian deduping and email threading so only the best copy is reviewed. Memorialize the agreed rules to avoid rework and later fights. Bring test results to the next meet-and-confer and lock in a tight protocol.
System triage helps focus on sources that hold real evidence and skip those that do not. Start with systems of record and key communication hubs tied to the claims and defenses. Defer low-utility sources like routine backups or noisy telemetry unless a strong link appears.
Document the reasons and preserve as needed so nothing important is lost. Revisit the triage as new facts come in to adjust quickly. Map the data universe and cut the low-value sources today.
Firm budgets with staged gates keep work in balance with the case stakes. Define phases such as scoping, sampling, core review, and targeted expansion. Set clear triggers to move to the next phase, like hit quality or issue coverage, and stop when returns drop.
Track spend and key metrics so leaders can make fast, informed calls. Use the plan to support proportionality arguments if scope creep appears. Define your phases and hard stops before the first collection.
Stratified sampling keeps discovery focused by testing small, well-chosen slices of the data. Break the universe into clear groups by source, date band, and role to reflect likely differences. Pull modest, random samples from each group and check how many items are truly relevant.
Use the results to raise or lower effort where the payoff looks strong. This approach gives data-driven proof of scope, which helps in court and in talks with the other side. Draft a short sampling plan and run a pilot this week.
Technology Assisted Review can stay proportional when training and validation are open and repeatable. Define a clear seed set, a holdout set, and rounds of training with simple success targets. Keep an audit trail that shows what was used, what changed, and why it changed.
Share validation summaries and error rates so trust grows on both sides. Use elusion tests to confirm that very few relevant items remain in the unreviewed pile. Stand up a TAR protocol with shared validation steps now.
