Essential Discovery
Essential Discovery - Attorney Careers


Deduping & Near Dupes

This white paper discusses best practices for addressing the presence of exact duplicate and near-duplicate (“near dupe”) documents within a dataset. Although the two are similar, practical applications for dealing with them are quite different. Deduping is best used as a tool for culling your dataset so that reviewers do not encounter the same document multiple times. Near duping, on the other hand, is a way to group similar documents to speed up the review. [Read More…]

Email Threading

As anyone who has reviewed long email chains understands, email conversations can get complicated. First, you have the original email; then there may be replies back and forth; the email may get forwarded to additional recipients; and, along the way, the chain may branch off into separate related (or even unrelated) side conversations. If documents are batched out chronologically by custodian, the individual emails that comprise a particular chain can be spread out among several different batches and different reviewers. As a result, the earlier emails are re reviewed every time a new email in the chain is encountered. This not only creates unnecessary duplication of efforts, but it also increases the risk of inconsistent tagging. Email threading offers a way to bring order to the chaos. With threading, emails are grouped by conversation so that they can be reviewed – and, ideally, batched – together. [Read More…]

Clustering and Categorization

Gathering the information in your database into meaningful sets is one of the best ways to prepare your team for an efficient and streamlined review. This white paper covers two options for grouping documents by related topic: categorization and clustering. Categorization is a user-informed way to group similar documents together. Categorization typically starts with having the review system generate an index of terms or concepts contained within the documents. The vendor, or in some cases, the project manager, can then feed in a small set of exemplar documents, which the system compares against the index to retrieve similar documents and group them into categories. Clustering, like categorization, is a way of sorting and grouping similar documents. Unlike categorization, which requires human reviewer input to inform the system’s output, clustering is usually an automated process run by the vendor. It is “automated” in the sense that there is no “seed set” of key documents that guides the system’s analysis of the database. The differences in how clustering and categorization work necessarily inform how the two tools are best used. [Read More…]

Predictive Coding

Although the eDiscovery industry has been abuzz with excitement over predictive coding, the truth remains that it is not yet in widespread use. This stems largely from a distrust of the technology – many still see predictive coding’s computations as a “black box.” The good news is that there are ways to use predictive coding that drastically increase the rate of review while still using the traditional approach of having attorneys review every document in the database. In this white paper, we explain how popular predictive coding tools work, suggest the best ways to incorporate these tools into your review workflow, address concerns about “black box” technology, and highlight important developments in case law that suggest a trend toward judicial approval for predictive coding. [Read More…]

Recommended Workflows

Review workflows are infinitely adaptable to meet the needs of your case. Included in this white paper are two sample workflows that show how technology assisted review (“TAR”) can be used to maximize review speed and accuracy. The workflows differ in the presumed level of comfort that the client has with using TAR. [Read More…]