Data Mining & Analytics
DOAR has successfully utilized data analytics to quickly identify irrelevant documents and remove them from the review stream thereby significantly speeding up the overall effort. Domain name isolation and classification allows for further refinement in both exclusion and inclusion of document groups.
Contextual relationships offer a different view into the document population and present patterns that may exist across the data set so you can start to see what relevance really represents early on. It will also group similar information together based on pre-determined content or some other statistical relevance which in turn lets you speed up the decision making process. Moreover, the application of this type of methodology lets you maintain consistency across the entire document review process and avoid discrepancies in the subjective coding of similar documents across different reviewers.
The use of analytics to find potentially privileged documents helps to ensure that the appropriate safeguards have been put into place by casting a wider net to identify those documents based on content and not just email address or the typical privileged language.
DOAR's application of these techniques not only allows attorneys to review documents that need to be produced faster, it also offers a distinct advantage at gaining insight into the production that you are receiving. For example, you can use documents that are most relevant to the case in your possession to quickly isolate other similar documents received from opposing counsel and determine to what extent they are in compliance with the document request.
Analytics Based Review
There are many circumstances that allow a party to completely bypass the manual document review process when utilizing data analytics. Several factors go into making this decision, one pair of which being the type of data that is available and the level of reliability that goes along with it. For example, while the use of analytics across scanned documents with OCR is far superior to standard keyword searching, the percentage of inaccuracies associated with the OCR process can sometimes throw off the analysis. On the other hand, certain types of native electronic data can be put through the conceptual categorization, issue coding and privilege segregation process with a high degree of accuracy and consistency and be put straight through to production without additional manual review. This methodology considerably lowers the cost of discovery, allowing corporations to perform much of the review in house while letting outside counsel focus on more substantive issues sooner.