IV. PROCESSING, PROTECTING, AND PRODUCING ESI 9. Processing, Filtering and Reviewing ESI a. Bag and Tag and Bag i. Store ESI on secured accessible platform for filtering. ii. Files tagged as: 1. Relevant 2. Privileged 3. Confidential 4. Discard 5. Damaged iii. Files bagged to meet production deadlines. b. Not organizing and indexing ESI can carry a sanction. c. Planning, Tagging and Bagging i. Define the outcome and what’s needed to achieve it. ii. Early Case Assessment iii. Process and index preserved ESI. iv. Filter the processed ESI 1. Filter through Reference Data Set of the National Software Reference Library – collection of digital signatures of known system files and software applications. v. Review and Analyze ESI 1. Redacting vi. Rolling Production? d. United States v. O’Keefe – Judge Facciola keyword search terms should be left to experts, not lawyers. e. Reducing ESI Volume i. Dupe identification ii. Finals versus drafts. iii. Near dupes from one of a kind. iv. Refer to creation, modified, or accessed dates. f. Trial and Error i. More importantly, the court expects that you use the trial-and-error approach and document your results with each trial. There is case law requiring a systematic approach. ii. FRCP 26(g) good faith effort required but efforts should be proportional to costs and benefits. iii. FRE 502(c) waiver protection only applies if you took reasonable measures in search iv. Objective searches based on dates and custodians. v. Subjective searches based on keywords and context g. Early Case Assessment i. Range of automation tools available ii. Vendors 1. Fios 2. Kazeon 3. Lexis Applied Discovery 4. vDiscovery 5. Access Data h. Processing ESI i. Standardize file formats ii. Document type specific data extraction 1. Zip files unzipped; email attachments. iii. Remove repetitive copies iv. Remove system files v. Inventory ESI vi. Extract metadata vii. Email thread detection viii. Export ESI to review database. ix. Create an index and word frequency list. i. Filters i. File size and type filtering – 1. Temporary files a. .tmp b. Begin with ~ or $ 2. Agreement with opposing side on certain file types. 3. Use forensic software that doesn’t rely on file extensions to find file types. 4. Custodian 5. Date Range j. Keyword or phrase searching i. Fuzzy searching. k. Deduping i. Within in each custodian’s ESI or within single repository. l. Concept Searching i. Uses the meaning for words or terms to find responsive documents. m. The Grimm Roadmap i. Victor Stanley v. Creative Pipe keyword search requires 1. Collaborative search – confer with opposing party to identify mutually agreeable search and retrieval method. 2. Best practices and data driven search a. Stay current with best practices. b. Prepare your defense – affidavits from experts attesting that reliable principles or methodology was used. c. Be cable of rapid course correction. n. Sampling i. Mandatory in large cases. ii. Black box approach won’t work – as per courts. iii. Take a statistically significant number of files. iv. Review files manually for ESI that you want it to identify or target. v. Use automated search methodology to mirror manual review. vi. Compare manual and automated results. vii. Reach a conclusion. viii. Judge Facciola and Grimm – perfection not the standard of discovery. FRE 702 expert standards make the process rational not perfect. ix. Get a clawback provision as hedge against sampling errors. o. Review i. Responsive ii. Irrelevant iii. Privilege iv. Private v. Work product vi. Redact vii. Corrupt viii. Encrypt ix. Password protected x. Hot xi. For further review p. Review Platforms i. Cloud computing – access services via the internet without knowing IT infrastructure. Pay for usage. ii. Software-as-a-Service (SaaS) hosted applications. iii. Dashboards – information portals that provide real time actionable information to track status, timelines, percent completion, cost estimates. iv. Drilldown to get additional details. q. How to Perform a Review i. Preserve relationships between documents or emails and their attachments. ii. Link duplicate files. iii. Preserve meta data in native file formats. iv. Verify tagging and redaction are consistent before disclosing.