Every user of Relativity knows that the tool offers an abundance of useful features. As a Relativity practitioner who often conducts demonstrations and training sessions on how to use the platform, it is often a fine art to cater a training session to a case team. A training session should demonstrate the important features of Relativity, while at the same time not making the training overwhelming by showing too much and distracting from the core review functionality.
Let’s be honest: 90% of the people who are attending training sessions will be reviewing documents and will not venture outside the core review interface. The remaining 10% may attend the initial training, but will be working in other aspects of the case in some form of case management. Scheduling more advanced training sessions is in everyone’s best interests but this often becomes a challenge due to the speed at which a case progresses and the workload associated with the case.
This two-part article will expand on basic training and will identify five key Relativity features which are often overlooked. These features are outside the platform’s more mainstream tools such as keyword searching and email threading. These features range from trial preparation, in the form of recording and reporting on important documents and events in your case, through to gaining a better understanding of your case by identifying key documents using conceptual analytics.
Let’s begin our exploration of Relativity features with the investigation stage of a matter. All too often, an investigation into a data set takes place too late in the case lifecycle to provide the maximum benefit. In an ideal world, documents should be processed and analysed as soon as possible in the case in to order to get a better understanding of the content. It is widely believed that agreeing on keywords during the Case Management Conference (CMC) is too early in the process to have a full understanding of the impact that these may have on the case. Agreeing on keywords without an in-depth knowledge of the case could result in an excessive number of documents, which are often not relevant. A poor choice of keywords could also result in missing key documents by excluding important terms, overlooking the use of code names and acronyms, or failing to recognise when the keywords are accidently spelt incorrectly in the document content.
The following Relativity features allow the legal team to leverage technologies to not only find key documents in the case quickly and easily, but allow more informed decision to be made. These informed decisions can in turn be used to kick start a review through prioritisation of documents.
Conceptual Searching allows Relativity to find documents that are conceptually similar to a string of text. Conceptual searching matches similar documents based on content, without the need for any of the words in the search to be present within the conceptually similar document.
It is commonplace that in the beginning of the case an internal document will be created to outline the key issues in the case, what has happened, and what reviewers should be looking for when reviewing documents. Using the text from this document (or a document of any length) a search can be created to find conceptually similar documents within the review workspace. During the search a minimum rank of conceptual similarity can be set, with the higher the rank resulting in identifying documents which are more similar in content to that of the submitted text.
As the internal document often contains a detailed overview of the key case details, it is common that interesting if not highly relevant documents will be identified by the system. These documents can then be coded and used with a variety of other Relativity features to help build a better understanding of your documents.
Find Similar Documents
Once relevant documents are identified, whether through conceptual searching, a traditional keyword search, or by utilising existing coding, Relativity can easily find similar documents in the case. The “Find Similar Documents” tool, is another form of conceptual searching. This method allows the user to request similar documents on the fly, rather than conducting a search on pre-existing textual content.
When the tool is used, conceptually similar documents are identified and provided in a list, sorted by the most conceptually similar documents first. This identification of similar documents can be conducted on the entirety of a document, or a selection of text within the document currently being reviewed.
Frequently, members of the legal team are provided a number of key documents by their client prior to conducting a document review. These documents can be quickly processed in Relativity in order to kick start the review. Using these processing documents, Relativity can be used to “Find Similar Documents” and will quickly locate important documents, in addition to those which the legal team are already aware of.
Clustering assigns documents into groups based on the conceptual content of the document, creating a choice of visualisations of a particular data set. Any documents that contain similar conceptual content will be grouped together by Relativity’s algorithms. These groups of documents contain sub-groups that allow the conceptual grouping to be more refined.
Cluster Visualisation allows these groups to be easily navigated and analysed by a user. Once documents have been coded in Relativity, cluster visualisation can then be used to identify clusters of interest (i.e. clusters which contain one or more documents coded to the desired coding). Whether this tagging is related to relevance, privilege, or issue, the coding can be overlaid as a “heat map” to easily identify clusters of higher importance. Using this method, a few documents can be identified by using a concept search, find similar documents, or traditional review and the clusters they belong to can then be batched for priority review.
Navigating cluster visualisation also allows users to locate nearby clusters (i.e. those that are not highly similar in conceptual content but are closely related). The distance (i.e. similarity) of the clusters can be adjusted to act as a ranking system for prioritisation. This allows users to identify similar clusters that may contain issues which were previously overlooked.
Although clustering and cluster visualisation are very strong tools, this process would not typically be used to structure the entirety of a review. These tools are best utilised as part of an investigation, to locate hot documents, and to prioritise documents so the most likely relevant documents are reviewed first. These tools can also be used to identify relevant documents in order to kick start a Continuous Active Learning (CAL) document review project.
Part two of this blog can be found here. If you’d like a direct experience, you can reach out to the Relativity experts at CDS to learn how you can use technology to enhance your eDiscovery and review needs.