Amongst all the articles which have been written on the emergence of predictive coding, aka Computer Assisted Review/Technology Assisted Review (CAR/TAR), it is noticeable that few contain references to any specific cases.
There are, of course, exceptions such as the case study Predicting the Future of e-Disclosure prepared by James Moeskops of Millnet in conjunction with Dominic Lacey and Jamie Tanner of Eversheds which has been featured here before and my own post Case Study: All in the mix [Smart e-Discovery, 26th June, 2012] in which Emma Kettleton described a project on which she was working earlier in the year.
Emma has now written further on the subject as the project has evolved and it seems appropriate to publish this update now.
The dispute concerned an allegation of overdrawing by one partner in a business and it soon emerged that there would be difficulties in agreeing appropriate keywords because much of the documentation, consisting largely of emails, concerned the ordinary conduct of the business by the parties. Finding appropriate keywords was a problem.
The solution adopted was CAR/TAR which resulted in a near 50% saving in cost for the client.
The updated case study can be found here: Case Study: All in the mix revisited