TAR, CAR, Predictive Coding, Integrated Analytics What Does It All Mean? Huron Legal provides advisory and business services to assist law departments.

TAR, CAR, Predictive Coding, Integrated Analytics

What Does It All Mean?

Huron Legal provides advisory and business services to assist law departments and law firms enhance organizational effectiveness and reduce legal spend. Huron Legal advises on and implements strategy, organizational design and development, outside counsel

management, operational efficiency, discovery solutions, and provides services relating to the management of matters, contracts, documents, records, digital evidence and e-discovery.

2

What is Predictive Coding?

“The use of software to automate the document review process in ediscovery based on artificial intelligence techniques.” Evan Koblentz, Law Technology News, March 2012

Better Name: Technology Assisted Review

© 2011 Huron Consulting Group. All rights reserved. Proprietary & Confidential. 3

• Using technology to assist with locating information responsive to discovery requests

• Requires a defined process and workflow that is documented and defensible

• “Seed sets,” data analytics, statistical sampling• Lawyers will be expected to understand, articulate, implement and

defend technology to identify relevant data

4

Brand New Technology? Not Exactly

What about• Keyword search• Clustering• Near duplicate identification• Concept search

• Other industries

Target’s Pregnancy Predictor--Real World Data Analytics

© 2013 Huron Consulting Group. All rights reserved. Proprietary & Confidential. 5

6

What Do I Need to Know About Technology Assisted Review?

• “Everyone” in the ediscovery world is talking about it• How it works• Cost implications• Not all methods are created equal• Developing case law

7

What Is Analytics?

© 2013 Huron Consulting Group. All rights reserved. Proprietary & Confidential.

8

Cases

• DaSilva Moore v. Publicis Groupe SA, 287 F.R.D. 182 (S.D.N.Y. 2012), adopted sub nom. Moore v. Publicis Groupe SA, 11 Civ. 1279 ALCAJP, 2012 WL 1446534 (S.D.N.Y. April 26, 2012)

• Global Aerospace, Inc. et al. v. Landow Aviation, L.P. , Consol. Case No. 61040( Va. Cir. Ct. Apr. 23, 2012)

• In re Actos (Pioglitazone) Products Liability Litigation, MDL No. 6:11-MD-2299 (W.D. La. July 27, 2012)

• Kleen Products, LLC, et al v. Packaging Corp. of America et al., Case No. 1:10-CV-05711 (N.D. Ill., Sept. 28, 2012)

The Current Approach to Document Analysis Is Not Sustainable

Data volumes continue to rise rapidly

Review workflows increasingly are not scalable to meet demand

Even with lower pricing, costs are not proportional to litigation risks

Most “predictive coding” solutions are fragmented, requiring merits counsel to fill in the gaps

Why TAR?

16© 2012 Huron Consulting Group. All rights reserved. Proprietary & Confidential.

Document

ReviewData Analytics ProductionHostingProcessingCollectionPreservation

Domestic and Offshore

Review Centers:

• 1500+ seat capacity in US,

Europe and India

Domestic and Offshore

Review Centers:

• 1500+ seat capacity in US,

Europe and India

Hard Copy & ESI

•TIFF responsive Docs

•Load file creation

Hard Copy & ESI

•TIFF responsive Docs

•Load file creation

Data Analytics

•Dedicated Analytics Center

•Multi-disciplinary team

•Search term analysis

•Defensible culling

Data Analytics

•Dedicated Analytics Center

•Multi-disciplinary team

•Search term analysis

•Defensible culling

17

70% of all eDiscovery

Costs

Document

Review

Volumes doubling

every 18-24 months

< 20% responsive rates,

even when using search

terms

REVIEW IS LESS ABOUT HIGH-LEVEL LEGAL ANALYSIS AND MORE ABOUT

DISCARDING IRRELEVANT DOCUMENTS

Ediscovery Process

PROJECT MANAGEMENT

© 2012 Huron Consulting Group. All rights reserved. Proprietary & Confidential.

eDISCOVERY CONSULTING

© 2011 Huron Consulting Group. All rights reserved. Proprietary & Confidential. 18

• The sampling population contains only one copy of each document and does not include documents without

concepts (e.g., pictures, Excel files, etc.)

• Each sample document selected is called an anchor

• Each document in the population gets aligned with an anchor document based on conceptual similarity

Integrated Analytics: “Anchor” Concept

© 2011 Huron Consulting Group. All rights reserved. Proprietary & Confidential. 19

Nearest Neighbor 1:

.75 SIM score

Nearest Neighbor 2:

.35 SIM score

IA: Similarity Scores

© 2011 Huron Consulting Group. All rights reserved. Proprietary & Confidential. 20

High SIM

score

IA: System Learning

Low SIM

score .25

.18

.28

.47

.35.33

.64.71.61

.95 .96

.87

.98

Goals:

1) Have little to no document “movement” when new anchors introduced

2) Have relatively high SIM scores throughout population

21

Integrated Analytics Workflow

© 2012 Huron Consulting Group. All rights reserved. Proprietary & Confidential.

22

Integrated Analytics

A Comprehensive, Integrated Process• Multi-Disciplinary Expert Team: Data Miners, Statisticians, Lawyers• Multi-tool approach (latent semantic indexing and machine learning)• Integrated process in a dedicated facility

Independent Statistical Verification of Process• Developed in collaboration with industry and academic experts• 99% confidence level (+/- 2.5% confidence interval)• “Accept on Zero” statistical validation

Expert/Defensibility Resources• Independent, dedicated resources (legal + academic experts)

Pricing• Pricing based on relevance percentage (< relevant = < price)• Capped Pricing: Places financial risk on Huron to properly cull out documents• Includes defensibility resources as part of pricing• Benchmarked cost savings of 30%