Electronic Document Conversion and Review Acceleration
ACT’s entire electronic discovery process revolves around providing quality services while limiting the number of documents that need to be reviewed by attorneys and the time attorneys spend with the documents they do review.
We call this Intelligent Document TriageSM and it can result in the savings of millions of dollars in complex discovery matters.
ACT Intelligent Document TriageSM—E-Discovery
ACT Multi-Tier ReviewSM—Tier 1SM
ACT begins the process of reducing the number of documents that attorneys must review—and the number of documents that must be processed—with a Tier 1, in-scope review.
A Tier 1 Review begins with a cataloging process that inventories every file on a media (hard drive, DVD, CD-ROM, etc.). From this point, ACT can provide complete audit trail as to the status and whereabouts of every file, from review through production.
The Tier 1 application directly reads the catalogued database. Users are presented with a Windows Explorer-like interface to review an index of documents, allowing them to browse folders at the highest levels and, at their discretion, delve deeper into subfolders and even view individual documents. These broad-based decisions eliminate (but preserve for contingencies) entire folders, documents and data that are obviously irrelevant, resulting in a reduction in data of up to 80%.
Our clients need not spend money to process this data or to have their legal team review the unnecessary documents one by one.
Electronic Document Conversion
ACT senior technical consultants, production engineers and project managers design unique project specifications based on the structure of input data and the desired output. Project-specific programs are then written to accommodate the needs of the project. Both production and project management review data and images for accuracy and to ensure all programs are running properly.
System File Removal
ACT uses industry methods for identifying the “signature” of a file. Through this signature, ACT can segregate files that are used by a computer’s operating system or program applications from those files that have been created by a user and might potentially contain evidentiary material.
Metadata Filter (Date, File Type)
ACT’s electronic discovery processes extract metadata programmatically. ACT can then apply a search filter to exclude documents that by description are obviously not relevant/responsive to a matter. This same technique can be used to prioritize documents as they go into review, focusing initial review energies on those documents that are most likely to be relevant.
De-Duplication
ACT calculates and assigns an “MD5 hash code” for each item. The use of the MD5 hashing algorithm is an industry standard and used in many security applications. Hash codes are calculated using the contents of an item, whether that item is an email or loose file. De-duplication can either be performed within a source (or custodian) or across sources.
Keyword Filtering
An aggressive process for making substantial reductions in a document population is the use of a keyword filter. Based upon a list of words or phrases (key case terminology and/or names of individuals or organizations), ACT can perform a query across some or all of the documents in an attempt to identify just those documents that are potentially relevant to a matter.
BucketizingSM
After Tier 1 review is completed, ACT takes the resulting database and creates a subset of the selected data. That subset is then processed according to the processing and review workflows designed for the case.
ACT’s document conversion process is designed to create efficiencies for the review of these remaining documents, actually accelerating the review process. Our metrics have shown us that reviewing documents the traditional way, by custodian, can be very inefficient. ACT has found that if it can organize subset populations of documents by content, then review speeds can increase from hundreds of pages per hour to thousands.
ACT refers to the process of content organizing as BucketizingSM. The chief idea behind BucketizingSM is understanding as much as possible about the collections and document requests to be able to sort the documents into like groups prior to review. These “buckets” are created with concept search tools which use text patterns and statistics—rather than key words—to determine a high probability that the documents in a particular bucket are related.
After BucketizingSM, the reviewer can focus on similar documents and make much quicker (and better) decisions. In some cases, subsets of very similar documents can be reviewed by batch or sampling, creating effective review rates of thousands of pages per hour.
ACT Multi-Tier ReviewSM—Native file
When time permits, Native File Review (reviewing documents in their original file format) is often the least expensive alternative for file review. Only responsive documents must be converted, obviating processing fees for all non-responsive documents.
ACT’s Native File Review features the extraction of metadata and text from each file into a database record, a control number assigned for audit control and preservation of work product purposes and the delivery of the database linked to the native files in the review tool of the client’s choice.
Attorney reviewers can review the native file alongside the related text and metadata, increasing review speed substantially.
ACT Multi-Tier ReviewSM—TIFF
When time is at a premium, ACT may recommend the documents be converted to TIFF image format prior to attorney review. Attorney document review is usually the slowest step in the discovery process, and often the quickest workflow is to TIFF all documents in advance so selected documents are ready for production immediately. Very high (anticipated) responsive rates for the collected documents, or subsets of documents, also point to a TIFF-based review.
At ACT, we always develop the best workflow for our clients, based upon their specific needs, expectations, budgets and timeframes. We never force a “one-size-fits-all” process.
Paper Document Conversion
Image Capture
Scanning services are provided by 31 Independent Certified Imaging Centers throughout the US. ACT controls the quality of the process by setting the imaging specification and then manages the project to ensure compliance. ACT’s custom quality control program is used to review all scanned images and corresponding cross-reference files prior to the coding process. This ensures that every scanned image is accounted for and that control number gaps, duplicates, overlaps and orphaned images or records are completely eliminated.
ACT Certified Imaging Centers
Objective Coding
The objective coding of scanned documents is performed at ACT’s coding facilities located in the Philippines. These two facilities are ISO 9000 certified underscoring our quality, commitment and focus on providing our clients with a superior product at a more reasonable price, while maintaining the confidentiality and security of their data.
Our coding teams utilize customized and project-specific coding programs developed at ACT’s Los Angeles data lab to ensure that case specifications are met. This also allows for document-type specific coding so that resources are not expended unnecessarily.
The coding module technology is based on the same quality control utilities we developed in the mid-1980s that resulted in ACT’s enviable industry reputation for quality excellence in the architectural integrity of coded data.
All fields for all documents are reviewed after initial phase of coding. In addition to this, prior to transmission to our Data Lab in Los Angeles, output files are submitted to a separate Quality Assurance department to conduct pre-transmission audit. No file will be transmitted unless it passes its required accuracy rating.
As the documents complete the coding process, the data is transmitted, via high-speed line, back to the Data Lab for final quality control, data formatting and mastering onto the appropriate media.
ACT ValidatedOCRSM
ACT’s OCR process is unique in that we track and report OCR quality statistics on three levels: for the entire project, at a document level and at a page level.
Each page is validated with a success ranking percentage and description (Good, Moderate, Suspect or Poor) for that page, which allows the legal team to focus on pages that may not lend themselves to a keyword search. OCR may be performed selectively by document type in order to save money.
Quality Control and Data Normalization
ACT initially earned its reputation in the litigation support industry in the 1980’s by writing programmatic utilities to analyze and fix scanned images & data produced by other suppliers. We continue to write and utilize custom utilities to evaluate and fix data, which has evolved, to the world of e-mail and other electronic documents. We believe our use of experienced developers, programmers and production engineers to write and implement programmatic solutions to solve complex data issues results in higher quality and lower costs for our clients as opposed to fixing data manually.
Quality Control of Scanned Images, Data & Cross Reference File
Whether ACT receives scanned images from one of our certified scan centers or receives opposing and third party scanned CDs, they all run through the proprietary ACTQC process.
ACTQC checks the cross reference file for accurate beginnings and endings of documents, control number gaps, page number gaps, orphaned images and references. We are looking for consistency in data. We run our proprietary structural, four corner check. Structural problems are the most common issues we come across: Format and inconsistencies. Once we identify the issues, we can provide our clients with solutions to solve them.
Normalization of E-mail and Electronic Files
With the advent of e-mail and electronic files, ACT’s QC process has evolved to examine electronic document populations. This process applies primarily to opposing and third party productions that we receive. There are a number of issues that ACT routinely tests for and offer solutions including:
- Architectural issues
- Normalizing fields
- Data mapping
Architectural issues
In our experience, every large production will have some architectural errors as the data gathered and processed, electronically or hardcopy, and reviewed by opposing counsel takes place over a long period of time. Generally, time and consistency of the processing is the culprit for errors as technical specifications change and modifications to the original process are changed (deliberately or accidentally). Additionally, data may come from multiple suppliers with differing specifications and processes. These occurrences can cause inconsistencies within opposing counsel’s data and productions.
ACT will test all images and database files against each other to assure opposing counsel has delivered architecturally sound data (meaning that all images that are referred to and all database records referred to be properly linked to each other and that all images referenced are in fact available and that all records have images). ACT considers these issues to be architectural data errors. Once evaluated, ACT will consult with our clients to discuss our findings and options including asking to have data resupplied, working with the data we have or fixing the data depending on how problematic it is.
Normalizing fields
ACT will also normalize fields of data so that the data incoming is usable and sortable. An example of this normalization would be Bates Number Normalization. In addition to providing the original bates number, ACT will, if necessary, create a bates number consistent across all productions so the Defense Group can appropriately sort the productions. For example, incoming productions may have bates numbers containing periods (AZ.00001.00001) or underscores (AZ_0000100001) or no punctuation, but spaces (AZ 00001 00001) or no spaces (AZ0000100001). These data discrepancies need to be addressed prior to loading this data onto a database to be used by the Defense Group. If ACT simply loads the Bates Number, as is, the Defense Group will not be able to sort the Bates Numbers appropriately. All of the instances above will disable sorting properly.
Data mapping
ACT considers data mapping to be part of the normalization process. Data mapping is the process of taking the incoming data fields and making sure all of the fields containing the same type of data are named the same. For example, the Author field may be called “From”, “Author”, “Creator”, “Authored”, etc. It is critical that all data is mapped appropriately prior to loading data. In this case, the Defense Group may decide to call the field “Author”. Mapping the data will insure the Defense Group is searching the appropriate fields and will reduce the number of total fields populated in the database.
’s Service Offerings