How Precise Scoping Reduces eDiscovery Cost and Risk

Many eDiscovery budgets grow before review begins. The cause is often not the number of truly relevant documents; it is the failure to separate five different decisions: what must be preserved, what must be collected, what must be processed, what must be reviewed, and what must be produced.

Those sets may overlap, but they do not have to be identical. Precise scoping protects the record while preventing unnecessary data from flowing into the most expensive stages.

Preservation is broader than active review

A legal hold may require preserving several accounts, devices, or systems because they could contain relevant information. That does not always require immediate forensic collection of every source, full processing of every container, or attorney review of every item. Counsel can preserve broadly while collecting and processing in phases.

The distinction should be documented. Identify what is preserved in place, what is collected, what remains deferred, and what event would trigger expansion.

Start with claims and factual questions

A collection request becomes more useful when it is tied to the disputed facts. Instead of “collect all communications,” identify the people, transaction, project, file, event, date range, or alleged conduct at issue. Then map those questions to likely data sources.

Useful scoping questions include:

• Which custodians are likely to possess unique information?
• Which systems hold authoritative records versus convenient copies?
• What date range is justified by the events and retention behavior?
• Which applications or channels were actually used?
• Are there known code names, project identifiers, domains, file paths, or data types?
• Does the review require complete conversations, file versions, or surrounding context?
• What metadata is needed to evaluate authenticity, sequence, access, or custody?

Map the downstream review before collecting

A collection method that looks efficient may create problems later. An export may flatten conversation context, omit attachments, change file names, exclude version history, or fail to preserve the metadata required by the review platform. Conversely, a full forensic image may capture vast volumes of irrelevant operating-system and personal data that counsel never needs to review.

Before collection, confirm who will host and review the data, what formats the platform accepts, which fields matter, whether native files or images are required, how messages should be represented, and what exceptions the review team can handle.

Use phased collections

Phasing allows the team to test assumptions before committing to the largest volume. A typical sequence might be:

1. Preserve all reasonably anticipated sources.
2. Collect priority custodians or authoritative systems.
3. Process inventories and high-level metadata.
4. Evaluate volume, date coverage, file types, channels, and obvious gaps.
5. Collect or process additional sources only when the first phase shows a need.

This approach can reveal that a cloud repository contains the authoritative record and makes a laptop collection less important, or that a small group of channels accounts for most responsive communications.

Apply filters only when they are technically and legally defensible

Date, custodian, folder, channel, account, file-type, and other filters can reduce volume. But filters should be tested against the source and the factual questions. A date filter may exclude a later edit or earlier creation event. A file-extension filter may miss renamed or embedded content. A keyword filter may miss images, code words, multilingual content, or messages that rely on context.

Document the filter, why it was selected, how it was applied, what it excludes, and any validation performed. Filters should be a reasoned method, not an invisible shortcut.

Control duplicate and system-generated data

Deduplication, threading, near-duplicate analysis, suppression of known system files, and exclusion of technically generated records can reduce review volume. Each decision affects context and must be aligned with the review strategy. Global deduplication may hide which custodian possessed a document; message threading may change how unique attachments are counted.

Specify whether custodian-level information, family relationships, conversation context, and source locations must be retained even when duplicate content is suppressed from active review.

Define a review-ready deliverable

A defensible handoff should identify the sources collected, date and method, filters, processing choices, fields, time handling, native and image strategy, text extraction, parent-child relationships, messages, exceptions, and quality checks. The review team should not have to reverse-engineer how the package was created.

The practical result

Precise scoping does not mean collecting too little. It means knowing why each source and transformation is included. The legal team retains the ability to expand while avoiding the automatic assumption that every preserved byte must become an attorney-review document.