At Relativity Fest I interviewed Cameron Tschannen of NightOwl Discovery about the use of email threading and textual near-duplicate identification in eDiscovery.
These tools, Cameron Tschannen said, are always used by default at NightOwl. They group documents together for organisational purposes and searches, and enable the safe removal of documents from review. This simultaneously reduces volume and increases quality.
Reviewers can see which documents are related, and can treat like documents in a similar fashion, whether for the review itself or for quality control analysis. For QC purposes, the tools make it easy to compare the treatment of documents within each group. It may be a reasonable assumption, for example, that if one document is privileged then they all are, but this is analysed in a second pass before being accepted as correct.
There is, inevitably, a cost to this, but the time and costs saved in the end come from the fact that the review goes faster, documents are coded more consistently, and the clients get a better product.
Cameron Tschannen says that clients are increasingly aware of the availability of these tools, and better informed about their use and value than they used to be.