How to Reduce Litigation Costs In the EDRM Continuum

The obvious question is: How does a company reduce discovery costs in a defensible manner? This article looks at some of the low-hanging fruit along the EDRM (Electronic Discovery Reference Model) continuum and provides some strategies for early decision-making to reduce overall litigation costs.

Information Management

Strong, audited and reinforced information governance (IG) policies are essential to reducing litigation costs. Implementation of an IG policy requires more than dusting off the document retention schedule that has not been consulted in years and that no one really knows how to put into practice. An effective IG policy is tailored to each business unit. A uniform company-wide retention period simply does not address the needs of departments required to keep certain data for specified periods under various regulations (e.g., Sarbanes-Oxley). Further, the use of personal computing devices, tablets and smart phones has diversified the locations where data resides and this must be addressed from an IG perspective.

IG policies should be reviewed annually with respect to how well they serve the company's business needs and to confirm that they comply with all applicable laws and regulations regarding document retention. A practice of regular review and revision will allow companies to proactively manage new storage mediums and update policies that are not working. Equally important is holding training sessions with different departments within the company so all groups understand where, when and how to retain or discard documents pursuant to the policy.

Developing a set of policies tailored to how and where employees work, and training them on those policies, will reduce the amount of data retained by the company. As a result, the amount of data subject to discovery will also be reduced.

Identification and Preservation

The well-established legal requirement that the duty to preserve is triggered when litigation is initiated or reasonably anticipated is no doubt familiar to everyone reading this article. The first step in implementing a litigation hold is to suspend any automated or manual deletion schedules. The second step is identifying employees that may have data relevant to the issues of the case or investigation. Department heads, key custodians and the legal department can assist with identifying potential custodians.

The third step is to give the potential custodians a hold memo and a simple questionnaire on the custodian's document retention and storage practices. (Pre-collection custodian interviews can be used instead of or as a supplement to questionnaires.) One purpose of the questionnaire is to identify where data is physically stored. Data locations may include personal devices like PCs and mobile phones as well as shared systems like databases, SharePoint repositories, cloud-based resources and network shares on a server. A questionnaire can also be a tool for the legal department to identify custodians who have not acknowledged the hold or who have questions about it.

The IT department is an important resource for making sense of individual custodians' document practices and locating potentially relevant data; for example, custodians will often use internal company or department naming conventions (e.g., “the X drive”), but do not know where the data is actually stored.

Collection

A documented custodian interview is the best way to identify potentially relevant documents and other files (and to eliminate clearly irrelevant files) outside of e-mail. The most efficient and defensible option is to have both litigation counsel and the vendor participate in the interview. It may be helpful for the custodian to have someone from IT present as well. The interview is an opportunity to ask the custodian substantive questions about the case and simultaneously identify specific file paths of electronic documents to be collected, describe any databases that relate to the litigation and identify any paper documents that may need to be collected and scanned. Although some data sources like databases require tailored collection methods, the vendor can typically perform a forensically sound targeted collection during the interview, which saves time, money and the internal IT department from having to perform collections. The interview is also a prime opportunity to identify potential keyword search terms along with names and other terms for the privilege search. Last, the interview is a chance to ferret out names of potential additional custodians.

With regard to e-mail, it is wise to be wary of e-mail collection tools. There are many tools that claim to allow a company to run keyword searches against its entire e-mail system in order to locate relevant e-mails. However, many have reported after evaluating these tools that there are issues searching .zip files, scanned documents like non-searchable .pdf files and less common file types, such as those generated by proprietary software programs. In addition, some of these tools do not supply a report stating which files were not searched. In most cases, the best practice is instead to have a vendor, in collaboration with the IT department, collect each custodian's entire mailbox. One exception to this rule is that many e-mail systems, including MS Exchange and Lotus Notes, support date restrictions in exporting mail data. For companies using such systems, outside counsel can work with the company to identify relevant time frames (typically with buy-in from opposing counsel) and the vendor can limit the collection accordingly.

The IT and legal departments should be kept in the loop during the collection process. IT staff can assist the vendor with the technical aspects of the collection, while the legal department can assist with scheduling and be available to answer custodians' questions.

Processing, Review, Analysis

Making a targeted collection based on custodian questionnaires and interviews, rather than a blanket collection from all employees or even just of all documents from identified custodians, is an effective means of reducing data volume at the front end of the discovery process. There are also sound techniques that can be utilized during processing and review to further cut down on volume.

Data Filtering

Automated data filtering during pre-review processing is a powerful means for reducing volume. The most commonly used filtering methods are de-NISTing and de-duplication. De-NISTing removes program files, malware and many other non-user created files by comparing the dataset to a list of known file types published and regularly updated by the National Institute of Standards and Technology, which gives the filtering method its name. De-duplication is an even more effective filtering means, especially in e-mail collections where there is a very high degree of duplication across custodians' mailboxes. Other filtering techniques that can be useful in particular cases are date restrictions and file type filtering. File type filtering is used to cull certain categories of files, like audio/video files or files generated by specialized software programs, that the legal team has determined are not relevant to the case. The vendor can generate a list of file types included in the dataset and the programs that created them, to assist counsel in making that determination.

Technology Assisted Review (TAR)

Technology Assisted Review (TAR) is one tool in the tool kit, and not a be-all, end-all solution. It is a proven means to quickly, and relatively inexpensively, hone in on the documents most likely to be responsive. There are several types of TAR, but one popular type consists of a subject matter expert (SME) reviewing a seed set of documents randomly generated by the TAR algorithm. The SME is typically an associate who is closely involved in the discovery process and knowledgeable about the production requests, such as the attorney who conducted the custodian interviews. After the SME determines which documents in the seed set are relevant, an algorithm is used to take that knowledge and apply it to the remaining documents in the dataset. Ratings are assigned to the documents based on likelihood of responsiveness.

This allows the legal team to prioritize documents for review, since “likely relevant” and “likely irrelevant” documents can be grouped into separate sets for review. Key members of the legal team are able to focus on reviewing the documents that are most likely to be relevant to the issues of the case, which is also invariably the smaller set, while cost-effective reviewers such as contract attorneys can be assigned to the likely irrelevant set.

Leveraging Keywords and Privilege Terms

During the review process, good communication between the vendor and legal team is critical to an effective review of relevant documents and QC of irrelevant documents. The legal team is responsible for identifying keywords and compiling a list of names and other terms for the privilege search. However, the vendor can and should assist the legal team in using Boolean operators, crafting complex search strings and integrating any available advanced review tools that might be helpful, such as relationship analysis in privilege searches.

In the context of TAR, the vendor can use keywords and privilege terms to run searches against both the likely responsive and likely non-responsive document sets. This permits the document groupings to be further refined for review. In the likely non-responsive set, review of documents predicted non-responsive but that had keyword hits can be prioritized over predicted non-responsive, non-keyword hit documents. Similarly with regard to privilege review, documents (and their families) that are predicted responsive and had privilege term hits can be prioritized for review and logging. Finally, the third most important group is the documents that were predicted responsive but that did not have keyword hits. Document-by-document review or advanced analytics like concept analysis can help the legal team discover new issues and related keywords for further searching.

Conclusion

By plucking the low-hanging fruit throughout the EDRM continuum (such as leveraging information management, collection and processing options for reducing data volume), legal departments can defensibly reduce the ever-growing costs associated with discovery. Companies can and should set important strategic imperatives that will help them address this problem now and in the future.

Ann Grayson is a partner in the Litigation Department of Barnes & Thornburg LLP's Indianapolis, IN, office. She also serves as chair of the firm's e-discovery, Data & Document Management Practice Group. Jim Norman is vice president of business development for Qdiscovery. The authors gratefully acknowledge the contribution of Raymond J. Biederman, an associate at Barnes & Thornburg, in the preparation of this article.

The amount of data a company generates grows with each passing day. It is important to develop strategies to reduce the amount of data subject to discovery obligations while staying current with legal and technology trends. A strong partnership with a vendor and law firm using sophisticated data review and collection techniques is essential to navigating the discovery minefield in a cost-effective way that is also defensible and fully documented.

The EDRM Continuum

It is no secret that discovery accounts for a large percentage of litigation costs. This is largely based on the volume of documents that need to be reviewed by a team of attorneys before making document productions to opposing counsel. The ratio of pages discovered to pages entered as exhibits is as high as 1000/1. See http://tinyurl.com/m9mnu8a. This means that the vast majority of documents companies pay to have reviewed never see the light of day after the initial review and production.

The obvious question is: How does a company reduce discovery costs in a defensible manner? This article looks at some of the low-hanging fruit along the EDRM (Electronic Discovery Reference Model) continuum and provides some strategies for early decision-making to reduce overall litigation costs.

Information Management

Strong, audited and reinforced information governance (IG) policies are essential to reducing litigation costs. Implementation of an IG policy requires more than dusting off the document retention schedule that has not been consulted in years and that no one really knows how to put into practice. An effective IG policy is tailored to each business unit. A uniform company-wide retention period simply does not address the needs of departments required to keep certain data for specified periods under various regulations (e.g., Sarbanes-Oxley). Further, the use of personal computing devices, tablets and smart phones has diversified the locations where data resides and this must be addressed from an IG perspective.

IG policies should be reviewed annually with respect to how well they serve the company's business needs and to confirm that they comply with all applicable laws and regulations regarding document retention. A practice of regular review and revision will allow companies to proactively manage new storage mediums and update policies that are not working. Equally important is holding training sessions with different departments within the company so all groups understand where, when and how to retain or discard documents pursuant to the policy.

Developing a set of policies tailored to how and where employees work, and training them on those policies, will reduce the amount of data retained by the company. As a result, the amount of data subject to discovery will also be reduced.

Identification and Preservation

The well-established legal requirement that the duty to preserve is triggered when litigation is initiated or reasonably anticipated is no doubt familiar to everyone reading this article. The first step in implementing a litigation hold is to suspend any automated or manual deletion schedules. The second step is identifying employees that may have data relevant to the issues of the case or investigation. Department heads, key custodians and the legal department can assist with identifying potential custodians.

The third step is to give the potential custodians a hold memo and a simple questionnaire on the custodian's document retention and storage practices. (Pre-collection custodian interviews can be used instead of or as a supplement to questionnaires.) One purpose of the questionnaire is to identify where data is physically stored. Data locations may include personal devices like PCs and mobile phones as well as shared systems like databases, SharePoint repositories, cloud-based resources and network shares on a server. A questionnaire can also be a tool for the legal department to identify custodians who have not acknowledged the hold or who have questions about it.

The IT department is an important resource for making sense of individual custodians' document practices and locating potentially relevant data; for example, custodians will often use internal company or department naming conventions (e.g., “the X drive”), but do not know where the data is actually stored.

Collection

A documented custodian interview is the best way to identify potentially relevant documents and other files (and to eliminate clearly irrelevant files) outside of e-mail. The most efficient and defensible option is to have both litigation counsel and the vendor participate in the interview. It may be helpful for the custodian to have someone from IT present as well. The interview is an opportunity to ask the custodian substantive questions about the case and simultaneously identify specific file paths of electronic documents to be collected, describe any databases that relate to the litigation and identify any paper documents that may need to be collected and scanned. Although some data sources like databases require tailored collection methods, the vendor can typically perform a forensically sound targeted collection during the interview, which saves time, money and the internal IT department from having to perform collections. The interview is also a prime opportunity to identify potential keyword search terms along with names and other terms for the privilege search. Last, the interview is a chance to ferret out names of potential additional custodians.

With regard to e-mail, it is wise to be wary of e-mail collection tools. There are many tools that claim to allow a company to run keyword searches against its entire e-mail system in order to locate relevant e-mails. However, many have reported after evaluating these tools that there are issues searching .zip files, scanned documents like non-searchable .pdf files and less common file types, such as those generated by proprietary software programs. In addition, some of these tools do not supply a report stating which files were not searched. In most cases, the best practice is instead to have a vendor, in collaboration with the IT department, collect each custodian's entire mailbox. One exception to this rule is that many e-mail systems, including MS Exchange and Lotus Notes, support date restrictions in exporting mail data. For companies using such systems, outside counsel can work with the company to identify relevant time frames (typically with buy-in from opposing counsel) and the vendor can limit the collection accordingly.

The IT and legal departments should be kept in the loop during the collection process. IT staff can assist the vendor with the technical aspects of the collection, while the legal department can assist with scheduling and be available to answer custodians' questions.

Processing, Review, Analysis

Making a targeted collection based on custodian questionnaires and interviews, rather than a blanket collection from all employees or even just of all documents from identified custodians, is an effective means of reducing data volume at the front end of the discovery process. There are also sound techniques that can be utilized during processing and review to further cut down on volume.

Data Filtering

Automated data filtering during pre-review processing is a powerful means for reducing volume. The most commonly used filtering methods are de-NISTing and de-duplication. De-NISTing removes program files, malware and many other non-user created files by comparing the dataset to a list of known file types published and regularly updated by the National Institute of Standards and Technology, which gives the filtering method its name. De-duplication is an even more effective filtering means, especially in e-mail collections where there is a very high degree of duplication across custodians' mailboxes. Other filtering techniques that can be useful in particular cases are date restrictions and file type filtering. File type filtering is used to cull certain categories of files, like audio/video files or files generated by specialized software programs, that the legal team has determined are not relevant to the case. The vendor can generate a list of file types included in the dataset and the programs that created them, to assist counsel in making that determination.

Technology Assisted Review (TAR)

Technology Assisted Review (TAR) is one tool in the tool kit, and not a be-all, end-all solution. It is a proven means to quickly, and relatively inexpensively, hone in on the documents most likely to be responsive. There are several types of TAR, but one popular type consists of a subject matter expert (SME) reviewing a seed set of documents randomly generated by the TAR algorithm. The SME is typically an associate who is closely involved in the discovery process and knowledgeable about the production requests, such as the attorney who conducted the custodian interviews. After the SME determines which documents in the seed set are relevant, an algorithm is used to take that knowledge and apply it to the remaining documents in the dataset. Ratings are assigned to the documents based on likelihood of responsiveness.

This allows the legal team to prioritize documents for review, since “likely relevant” and “likely irrelevant” documents can be grouped into separate sets for review. Key members of the legal team are able to focus on reviewing the documents that are most likely to be relevant to the issues of the case, which is also invariably the smaller set, while cost-effective reviewers such as contract attorneys can be assigned to the likely irrelevant set.

Leveraging Keywords and Privilege Terms

During the review process, good communication between the vendor and legal team is critical to an effective review of relevant documents and QC of irrelevant documents. The legal team is responsible for identifying keywords and compiling a list of names and other terms for the privilege search. However, the vendor can and should assist the legal team in using Boolean operators, crafting complex search strings and integrating any available advanced review tools that might be helpful, such as relationship analysis in privilege searches.

In the context of TAR, the vendor can use keywords and privilege terms to run searches against both the likely responsive and likely non-responsive document sets. This permits the document groupings to be further refined for review. In the likely non-responsive set, review of documents predicted non-responsive but that had keyword hits can be prioritized over predicted non-responsive, non-keyword hit documents. Similarly with regard to privilege review, documents (and their families) that are predicted responsive and had privilege term hits can be prioritized for review and logging. Finally, the third most important group is the documents that were predicted responsive but that did not have keyword hits. Document-by-document review or advanced analytics like concept analysis can help the legal team discover new issues and related keywords for further searching.

Conclusion

By plucking the low-hanging fruit throughout the EDRM continuum (such as leveraging information management, collection and processing options for reducing data volume), legal departments can defensibly reduce the ever-growing costs associated with discovery. Companies can and should set important strategic imperatives that will help them address this problem now and in the future.

Ann Grayson is a partner in the Litigation Department of Barnes & Thornburg LLP's Indianapolis, IN, office. She also serves as chair of the firm's e-discovery, Data & Document Management Practice Group. Jim Norman is vice president of business development for Qdiscovery. The authors gratefully acknowledge the contribution of Raymond J. Biederman, an associate at Barnes & Thornburg, in the preparation of this article.