Procter & Gamble ediscovery Practices www.encase.com/ceic Topics 1. Why Procter & Gamble (A Fortune 50 company) conducts ediscovery and Forensics with reliance on Encase Enterprise v7 and ediscovery v5. 2. Case study: P&G revamped its ediscovery model which resulted 4MM savings in 8 months and reduced 4-6 week ediscovery case startup to 2 days with 0 defects. 3. How to make your ediscovery teams become relevant as versus being marginalized or be seen as just another "Information Security" group. 4. Off-topic discussion points including dealing with vendors whose interest may not align with yours, BYOM / BYOD and making sure appropriate ediscovery controls are in place for future applications. Page 2 Scott VanNice, Procter & Gamble 1
TOPIC Why Procter & Gamble (A Fortune 50 company) conducts ediscovery and Forensics with reliance on Encase Enterprise v7 and ediscovery v5 Page 3 Procter & Gamble ediscovery Stats why we rely on ENCASE YEAR CASE LOAD CUSTODIANS Aprox GB (assume 7 GB each custodian) 2009 4 159 1113(1.1TB) 2010 21 170 1190 (1.2TB) 2011 29 223 1561 (1.5 TB) 2012 21 194 1358 (1.4 TB) 2013 6 120 1595 (1.6 TB) Aug2014 now 10 47 300 GB (difference is a new vendor) Page 4 Scott VanNice, Procter & Gamble 2
Procter & Gamble ediscovery Stats why we rely on ENCASE 69% of the projects are less than 5 custodians 14% of the projects are between 6 and 15 custodians 8% of the projects are between 16 and 30 custodians <4% of the projects are between 30 and 50 custodians <4% of the projects are over 50 custodians 90% of my ediscovery projects end within a year Page 5 TOPIC Case study: P&G revamped its ediscovery model which resulted 4MM savings in 8 months and reduced 4-6 week ediscovery case startup to 2 days with 0 defects Page 6 Scott VanNice, Procter & Gamble 3
P&G IN-HOUSE MULTIPLE VENDORS Page 7 TODAY P&G Counsel Review Services Specialty Services Platform Platform Platform No scale Black Box with little control Higher cost Inefficient workflow TOMORROW P&G Platform Counsel Review Services Specialty Services Leverage scale Transparency and control Cost predictability Defensible workflow 8 Scott VanNice, Procter & Gamble 4
InsideCounsel puts cost at 3.5M per GB. Minnesota Journal Of Law, Science and Technology puts average cost per GB at 5M to 30M. RAND Institute ( Where My Money Goes ) estimates it can be as high as 18M per GB for review. All argue the same elements: 1. It is unrealistic to consider increasing human labor for review to reduce costs. The study endorses computer assisted technology including predictive coding. 2. Recommendations includes (1) improve tracking of costs of production and preservation and (2) identify vendors who do invest and leverage facilitating predictive coding and computer categorization strategies e.g. grouping detection. 3. Suggests at least two other options with predictive coding such as: (1) use it as a qualityenhancement process where small number of documents with the highest scores are sent to attorneys most connected to litigation and those with lower scores might be sent to lower-cost contract attorneys; or (2) enhance efficiency by have eyes-only review on only those documents with highest scores and offer those to opposing counsel as part of rolling production schedule and negotiate any need to review the remainder. Scott VanNice, Procter & Gamble 5
ISSUE: ediscovery is both a legal and a business process with overall goal of controlling total cost of ediscovery without losing defensibility. CURRENT FACTORS AS OF NOW: Aggressive data growth Too fast and too big Insourcing vs outsourcing of ediscovery activities Too many people; too many bottlenecks; too many duplication of effort. Need to find a balance. No economics of scale Forcing vendors to compete for a small step in the EDRM process creates little incentive to consider overall cost or protect chain of custody process. Globalization and new technology Identify where to play due to increases in ediscovery requests both inside and outside of the States and new changes in searching (e.g. predictive coding). KEY MESSAGE HERE: Consolidate the hosting, review and production processes into 1 or 2 vendors to drive standardized processes and leverage economics of scale. BOTTOM LINE: Clearly a competitive market with vendors offering nearly the same products / service which is why vendors depend on the following strategies: Charge by hour / GB / Page is strategy creating an illusionary sense of savings. Shift back and forth from gigabyte model or document pricing model. Encourage rolling uploads so they can charge more for hosting and project management hours additional processing, re-filtering, or new uploads from custodians. Taking advantage of the fact that most clients (law firms or in-house counsel) pass the costs through to others Keep their invoices ambiguous and then offer a 20% reduction if contested Suggest savings through predictive coding, flat rates, new technology then try to recoup in other areas (site activation, actual processing by GB, or bill hourly by tech support, etc) Scott VanNice, Procter & Gamble 6
Objective: How much does it approximately cost for the vendor to ingest (load), set up, review, produce, and host each gigabyte (GB)? Roadmap: No fuzzy math. Focus on vendor quotes and work backwards. Make it objective. Be conservative. Use results to identify unit of measurement. 1. Focus on post-p&gcollection and cull. 2. Align on basic assumptions ( vanilla ediscovery services ) 3. Pull SOWsfrom 5-7 vendors from current year. As vendors do not all adhere to standard units of measure, do an apples to apples comparison against: pre-process, loading, technical review, production, hosting, and management 4. Identify low end to high end cost using this vendor comparison. 5. Average low end and high costs identified to determine the cost per gigabyte. This is our unit of measurement. 6. Extrapolate the data against your case load index 7. Take findings and do a check with the industry and see if it holds up.. Transparency, Predictability and Control Keep overspend or underspend to a close margin of error Keep variables to a minimum and minimize risk disruption MODEL WHAT? DISCUSSION ANNUAL FLAT FEE Pay X for full year to be invoiced monthly. Model is simple. Penaltyif cap is reached (i.e. pay more). Risk is real. Difficult to charge by BU DATA SUBSCRIPTION PLAN You payx amount of blocks (hours, terabytes, databases, users) Modelcan get convoluted. Penaltyif cap is reached (i.e. pay more). Risk is real. Dfficultto charge by BU FLAT FEE PER GB You pay x amount per GB. Easymodel to understand This benefits some more than others i.e. not always fair Less confusion with BU charging TIERED BUNDLING Youpay x amount per tier. Pick and Choose. You get exactly what you want. Easy model to understand Less confusion with BU charging Fairness is evident for all Scott VanNice, Procter & Gamble 7
LONG TERM GAIN and OPTIMIZATION FUTURE BENEFITS SOFT BENEFITS Moving active data from other vendors at lower cost* One stop shopping center (e.g. managed review services) Leverage expert advice and share Security consolidation Efficiency with law firms and third parties Less outside counsel billable hours Proactively position P&Gwith future trends (legal + IT) First Consumer Sensing (confirm need and scope) Missing necessities (automatic out or no wow?) Cull vendors final list (6?) invite for deep dive Attorney scores (QFD) Industry ratings / review Second Consumer Sensing (debriefing and confirm must have v. nice to have services) External benchmarking / Client checks Fee cost model evaluation and selection w/ Legal or Purchasin Pricing and scoring Compare and Contrast Third Consumer Sensing Consensus on recommendation, governance, and savings Scott VanNice, Procter & Gamble 8
TOPIC How to make your ediscovery teams become relevant as versus being marginalized or be seen as just another "Information Security" group 17 Know your path and end objective Page 18 Scott VanNice, Procter & Gamble 9
SIMPLIFY DECISION MAKING: ediscovery Board (Governance) P (Process Leader) A (Approver) C (Contributor) E (Execute) Page 19 Governance YEAR ONE This PACE model launched P&G ediscovery Services (2007-2008) Case Management (Ongoing) MEASURE AND SCORE RESULTS 2/3 SAVINGS ediscovery Board FTI 2/3 START-UP TIME In-house counsel ediscovery PM Outside counsel TRANSPARENC Y REPEATABLE + DEFENSIBLE Lit Director(A) ediscovery Services (P/E) Parties from various disciplines (C) Third Parties (e.g. COC) USER SATISFACTION CURRENT to FUTURE TRANSITION P = Process Leader A = Approver C = Contributor E = Executor Scott VanNice, Procter & Gamble 10
Establish a ediscovery Policy Page 21 Simplify.. Simplify.. 80/20 of your work Page 22 Scott VanNice, Procter & Gamble 11
Be an interpreter: Corporate v Legal Outsourcing would cost $250 per hour There is no 9-5 work schedule. With 2 headcount, you can save 120M-150M a year per person. DRIVING IT HOME TO CORPORATE IN A SNAPSHOT Aggressive data growth Too fast and too big Insourcing vsoutsourcing of ediscovery activities Too many people; too many bottlenecks; too many duplication of effort. Need to find a balance. No economics of scale Forcing vendors to compete for a small step in the EDRMprocess creates little incentive to consider overall cost or protect chain of custody process. Globalization and new technology Identify where to play due to increases in ediscovery requests both inside and outside of the States and new changes in searching (e.g. predictive coding). KEY MESSAGE HERE: Drive standardized processes and leverage economics of scale. Page 23 MEASURE AND SHOW IMPROVEMENT Suggest you develop a single scorecard, under 4 rubrics: Technology, Education, Network Partnerships & Capability. Get alignment with management and Legal Get consensus on immediate next steps: Commit to reassessment in next FY Page 24 Scott VanNice, Procter & Gamble 12
Master Title TOPIC Additional discussion points such as dealing with vendors whose interest may not align with yours, BYOM / BYOD and making sure appropriate ediscovery controls are in place for future applications This is to be a free-flowing discussion Page 25 Scott VanNice, Procter & Gamble 13