Ride the Lightning
Cybersecurity and Future of Law Practice Blog
by Sharon D. Nelson Esq., President of Sensei Enterprises, Inc.
A Review of Two Predictive Coding “Dummies®” Books
December 4, 2012
Several weeks ago, I asked for volunteers to review two "Predictive Coding for Dummies" books, one by Clearwell (Symantec) and one by Recommind. Two brave souls stepped forward. The first was Terry Dexter, thanks Terry, for the time you obviously invested. I hasten to add (since several people asked) that I am no relation to Matthew Nelson, the author of one of the books! In fact, when we hit Ellis Island, our name was Nielsen – like so many other Swedes, the bureaucrats instantly changed the name to Nelson. The true names of my dad and grandparents are inscribed on the Immigrant's Wall of Honor on Ellis Island. And now to the reviews . . .
************************************
Predictive Coding For Dummies®, Symantec(TM) Special Edition by Matthew D. Nelson, Esq. Copyright © 2012 from John Wiley & Sons, Inc. 111 River St. Hoboken, NJ 07030-5774 ISBN 978-1-118-48198-1 (pbk); ISBN 978-1-118-48237-7 (ebk)
Predictive Coding For Dummies®, Recommind Special Edition author(s) not listed, Copyright © 2013 from John Wiley & Sons, Inc. 111 River St.Hoboken, NJ 07030-5774 ISBN 978-1-118-52167-0 (pbk); ISBN 978-1-118-52230-1 (ebk)
Not being known as someone who won't accept a challenge, I read both books cover to cover (several times). In full disclosure, I am not an attorney (or played one on TV); I am simply a techno-geek with a Bachelor of Arts in English and strong interest in the tools, techniques and methods involved with electronic discovery (eDis). This review is based upon my reading and understanding of Predictive Coding, which, in turn, is based upon a combination of 30 years in Information Science & Technology and extensive research into the wild wooly world of electronic discovery. Any and all comments are mine and not that of Sharon Nelson (the individual) or Sensei Enterprises, Inc.
Up first: Predictive Coding For Dummies®, Symantec(TM) Special Edition by Matthew D. Nelson, Esq.
My initial impression of this book was good. The format follows the standard “Dummies” format and structure while legal and technical concepts are presented in a clear, easily understood manner. Nelson's writing flows from one paragraph to another and doesn't introduce new terms without first explaining them. The reader is immediately informed as to the what and why of electronic discovery. From the third paragraph onward, the reader is gradually immersed into a sometimes murky world.
This excerpt from the Introduction sets the tone:
“Predictive coding technology is a new approach to attorney document review that can be used to help legal teams significantly reduce the time and cost of eDiscovery. Despite the promise of predictive coding technology, the technology is relatively new to the legal field, and significant confusion about the proper use of these tools is pervasive. This book helps eliminate that confusion by providing a wealth of information about predictive coding technology, related terminology, and the proper use of these tools.”
Specific comments:
Beyond the excellent writing, this book contains many positives and negatives; some of which I present here.
Positives:
- The cost in terms of timeliness, accuracy and productivity is compared to manual review.
- Nelson introduces the Electronic Discovery Reference Model (EDRM) within the first three (3) pages. The subsequent discussions regarding potential costs is emphasized by illustrating the enormity of the potential volume of Electronically Stored Information (ESI). This early introduction is also valuable when process defensibility is introduced.
- The concepts of sanctions, privileged information, human v machine reading/review and risk are easily distinguished. Again, the “whys” for such concepts, easily understandable to a First Year Law Student, are easily understood for the layperson.
- The inclusion of website addresses to provide additional information is most welcome. Indeed, references to a predictive coding cost estimate page and to a Ralph Losey article helped me gain a deeper understanding of the planning and execution of a PC effort.
- A separate step in Nelson's work flow considers Privileged Information. While no one on either side of a litigation struggle want to divulge such data, it can and does happen. Predictive Coding is not presented as a palliative cure-all for such 'ooopsies'; however the book does go far in helping the reader comprehend the necessity of conducting separate actions to reduce if not eliminate the probability of such an event occurring.
- The three prominent eDis cases (DaSilva-Moore, Kleen Products and Global Aerospace) are discussed relative to First Generation PC tools and Judicial Guidance.
Negatives:
- Clearly, this book is written and produced to influence litigators and law firms to orient themselves towards Symantec and Clearwell. Hints are subtly placed throughout the book. While not explicitly mentioning any names, the implication is clear and gets more obvious starting at Chapter 6. A more neutral, objective content makes more sense to someone who is already familiar with the eDis process.
- There is no discussion on the difficulties of using Optical Character Recognition (OCR) or different character set based ESI. All data is presumed to be 100% compatible and ANSI compliant.
Recommendation:
This is an excellent book to give to clients, new litigation support personnel, paralegals, etc. involved in the beginnings of any litigation where the use of Electronic Discovery tools is likely.
Next up: Predictive Coding For Dummies®, Recommind Special Edition author(s) not listed
My initial impression was guarded. The format follows the standard “Dummies” format and structure but the content reads like someone mashed several marketing 'White Papers' together. This impression is further supported when comparing copyright dates with the Symantec book. Indeed, a stark comparison between these tomes is like comparing apples to oranges.
Positives:
- It's short.
Negatives:
- Only nine (9) pages (25%) have any direct relationship with the subject matter. Twenty-eight (28) pages (~77%) are more closely related to marketing collateral. The very topic of Predictive Coding is introduced to the reader at page 11!
- The reader is constantly bombarded with the cost differential between manual and automated document review. Figure 1-2 in this book compares savings in 3 types of cases (IP, Second Request & Tort). Linear (Manual) Review is compared to Predictive Coding and, of course, PC wins every time. However there is no mention as to the style of the PC effort (and related costs) – were documents reviewed in house or by a services provider?
- There is zero mention of risk, sanctions or privileged information. In fact, a reader may develop the idea that any Predictive Coding tool takes care of any such occurrences.
- There is no discussion on the difficulties of using Optical Character Recognition (OCR) or different character set based ESI. All ESI is presumed to be 100% compatible and ANSI compl
iant. - Wh
at are 'Frankenstacks'? This book is supposed to help IT Managers who already understand the hurdles of application incompatibility. - The book is very difficult to read. The workflow discussion does not follow the accompanying diagram (Figure 2-1) and even introduces the concept of 'Predictive Analysis' without any further discussion.
- The book makes blatant reference to Recommind's product. Indeed the content of the entire document builds to the conclusion that only Recommind has the capability to successfully conduct electronic discovery.
Recommendation:
This is a very poorly written book using a style that insults the reader's intelligence. A cursory Bing or Google search would a better investment in time and money.
That's it for this review. Both books are available for download over the Internet. If you'd like, take a look for yourselves.
Listen to the Podcast
