0.0 Introduction

Download the complete paper: EDRM Guidelines for E-Discovery Processing v3CC4

E-discovery[1] involves the exchange, analysis and review of electronic files, email and other information stored on a computing device.[2] Over the past two decades, it has become a common part of U.S. litigation and regulatory investigations, spurring the growth of a multi-billion dollar litigation support market.[3] It has also spawned a new specialty for legal professionals, many of whom now focus their practices on, and get certified for, this aspect of litigation.

The goal of an e-discovery effort is to surface relevant information for trial, arbitration or a hearing. The process begins with the identification of electronically-stored information (“ESI”) that may be relevant to the matter and ends with the production of responsive, non-privileged ESI to a requesting party. In between are a series of steps designed to move data from identification and collection through processing, review, analysis, and production.

The stages through which ESI moves from its original location through trial or a hearing are depicted in the EDRM (Electronic Discovery Reference Model), which is a widely-accepted conceptual model of the e-discovery process.

Over the years, the model has evolved and is often used to help people understand how the ediscovery process works.


E-Discovery Processing

These guidelines, adapted from Craig Ball’s seminal work, “Processing in E-Discovery, A Primer, ©2019, will focus on the processing stage of the EDRM. As you can see from the EDRM model, processing bridges the gap between preservation/collection and review/analysis.

During processing, ESI that has been collected using a variety of methods and tools is processed through a series of steps  and prepared for review. In the early days, processing included printing ESI for manual review. As ESI volumes increased, printing to paper no longer made sense. Modern processing software converts ESI into a standard format for electronic review, with the resulting data often being loaded into specialized litigation support software.

Here is one look at the complexities underlying a standard processing workflow from Craig Ball’s Processing in E-Discovery primer.[4]

In sharing this diagram, we are not suggesting that it provides the only order of steps to perform this work (e.g. OCR is usually done after you filter (DeNIST, De-Dupe, Date), in order to reduce the amount of files it needs to run through). There are many ways to attack the problem. Rather, this is a visual view of processing, one that may help readers better appreciate the complexities underlying this stage of the EDRM.



These guidelines will help elucidate this often overlooked stage of the EDRM and will help readers better understand the basic steps involved in processing and the needs of the legal professions at this stage. In that regard, these guidelines are addressed to those who (a) use processing products, (b) support these products,(c) processing products, and (d) are new to e-discovery and want to learn about the processing stage.



In preparing the guidelines, the drafting team worked from the following principles:

  1. Our focus will be on processing functions rather than processing products. It will not discuss the individual strengths and weaknesses of specific products.
  2. Our intent is to describe the processing functions required for successful e-discovery processing and to note functions deemed optional, albeit desirable, in some instances. Our goal is to set a floor, not a ceiling.
  3. Processing consists of a number of tasks. A single product may attempt to accomplish all of these tasks or only a subset.
  4. A processing product may also attempt to facilitate other functions, such as data review. Similarly, one may rely on a single product for processing or several products to achieve that aim.
  5. As a cardinal rule, processing operations should not alter or destroy files or metadata.



Like the EDRM itself, processing functions proceed through a number of logical steps which, for convenience, are referred to as “phases.”  Following a lawyer’s penchant for chronology, the phases are ordered to follow the ESI from the point where collection is completed to when processed data is ready for review.

This is not to suggest that a good processing engine must work in this order or that any product must perform every step in each phase. Rather, the goal is to present the key processing steps in a logical order to encompass all functions traditionally included in the processing stage of e-discovery.

The key phases for the processing stage of the EDRM are:

  • 0 ESI Ingestion and File Extraction
  • 0 Initial Filtering
  • 0 Text, Metadata and Image Extraction
  • 0 Output
  • 0 Reporting

These guidelines will discuss each phase and how they may be used to meet legal industry needs.


[1] The term is shorthand for electronic discovery and is sometimes spelled eDiscovery or ediscovery. In the U.K. and some other jurisdictions, this process is known as e-disclosure. For convenience, we will use e-discovery to refer to all of its variants.

[2] See EDRM Glossary at 1106 ( https://edrm.net/resources/glossaries/glossary).

[3] E-discovery is less common outside the U.S. Many civil law countries, including those in the EU, do not allow e-discovery or limit it to evidence needed to support a claim or defense at trial. In contrast, most common-law jurisdictions, such as the U.K., Australia or Canada, allow e-discovery but often with a more limited scope than in the U.S.

[4] For more about ESI processing read Craig Ball’s “Processing in E-Discovery, A Primer,” published in 2019 and available here. It is an excellent source for those seeking to better understand this stage of e-discovery and to learn about industry requirements for processing functionality.