The World Is Flat: Handling Foreign Language Documents in eDiscovery Projects

Early in my (Darryl Shetterly [the author]) eDiscovery career I learned the unanticipated presence of foreign language documents can throw a project budget and timeline for a loop.  We were on budget and ahead of schedule when . . . BAM!  Lots of Asian language documents!  The English speaking team I had so carefully built and trained was suddenly unable to complete the task.  What to do?

The world is flatter now.  Anecdotally, we encounter an even higher volume of foreign language documents today as we help our clients comply with their eDiscovery obligations.  Foreign language documents introduce an additional layer of complexity and without the proper mix of the right people, process and technology, costs rise and deadlines slip.

Here are a few things to consider when facing foreign language documents in eDiscovery:

 

PEOPLE

Much ink has been spilled over communication in the eDiscovery context: legal and IT notoriously speak different languages and many litigants seem to speak different languages when negotiating eDiscovery issues.  When you introduce German, Italian or Spanish documents into the mix, the “people” element can get downright unwieldy.

Project Management:  Ever tried to manage a team of French speaking attorneys as they review documents written in French?  Answering their questions about how to categorize borderline documents and checking their work for quality presents challenges.  On a project of any size, you need a project manager (or team lead) that speaks the language and understands your case.

Human Translation v. Machine Translation:  Human translation is more accurate; machine translation is cheaper.  Machine translation is often called “gist translation” because it gives you the gist of a document.  There is a time for each.  Given cost and quality constraints the workflow model typically is:

1. use a native speaking review team to categorize documents identified for review and get a certified human translation of the subset of documents that are most relevant;  
2. use machine translation to allow an English speaking review team to categorize documents and get a certified human translation of the subset of documents that are most relevant; or 
3. use a hybrid of these two methods based on triaging the larger document population by content, e.g. by custodian or document topic identified by clustering or other technology tools. 

Remember, not all human translation or machine translation software is created equal; you often get what you pay for.

Custodians:  Setting aside time zone issues and complex privacy protocols, interviewing a document custodian in a country on the other side of the world and collecting their relevant documents are tough tasks.  The custodian may not speak your language, or worse, may misunderstand your request and point you to the wrong documents.  Use a translator to facilitate the interview or have the interview questions translated.  In projects involving a high volume of custodians, it may be more cost-effective to train a translator to perform the interviews.

Review Team:  If you need 40 attorneys on short notice to categorize Russian language documents in nearly any medium sized city in this country you will pay a very high rate – if you can even find them.  Consider using technology to tier documents by importance and route the lower priority documents to a non-attorney review team.

Cost: Multilingual attorneys are often significantly more expensive then their English-speaking equivalents.  This is another good reason to leverage a non-attorney review team if it makes sense for your case.

 

PROCESS

Noodle through the impact foreign language documents will have on your current workflow and establish processes to mitigate this impact.

Workflow:  Categorizing documents by language and routing them to the team identified to review documents in that language is harder than it sounds.  Your Spanish language review team will need a seamless and automated method to route mischaracterized English documents to your English review team.  Add a tag to the coding form for each language and automate the process of queuing documents for review by the correct team. Why bother, you ask, can’t the Spanish review team speak English too?  Cost – remember, you are paying the Spanish team more so you want to have your English team handle as much as they can. 

Communication Protocol:  Whenever you have multiple teams working on a project there is a potential for communication breakdown.  Establish clean lines of communication so each review team is getting the same information and categorizing documents consistently.  Since information needs to flow in both directions, identify the person on the litigation team that should be notified when one of the teams locates a “hot” document.  Also specify how that notification should occur.

Document Production:  Producing foreign language documents may be more expensive and time consuming than their English language equivalents.  The cost and time required to translate an entire document production can be considerable.  Some government agencies require documents produced in response to a government investigation be produced in English or with an accompanying translation.  Discuss production language in your 26(f) conference and incorporate that decision into your eDiscovery protocol.

 

TECHNOLOGY

Once you have the right people and process, you still need technology that can handle your workflow.
   
Review Tool Functionality:  In addition to the native documents and corresponding text loaded at the outset of the project, you may also want to load a machine translation version and/or a human translation version.  If you add an image (for drawing redactions) and a production image for deposition and trial preparation, you will have four or five versions of a document in your database.  How will your review tool group these versions?  How much will it cost to host these additional versions?

Global Delivery Platform:  If you have trouble staffing a review due to the lack of attorneys in your area that speak a specific language, then you need to take the work to the attorneys.  To take your work to a different city or country you will need access to office space, computers, keyboards, sufficient bandwidth to support a review team and building access restrictions to avoid a data security breach.   

Search Functionality:  Clearly the English language search term list you negotiated during the 26(f) conference will not retrieve documents in other languages.  There are technologies that will translate your search terms to other languages, but these tools typically depend on strict word translations and may not account for cultural differences in word usage.  Ask the custodians that created the foreign language documents what terms they would use to identify the relevant documents.  Work with opposing counsel to develop a list of search terms that make sense.

Language Identification: 
When using technology to group documents by language, it is not uncommon for 20 percent of the documents to be mischaracterized.  Mischaracterization typically occurs because the machine picked up on a name, address or phrase in an English document that it identifies as non-English, e.g. Carlos Rodriquez or custodians who use non-English phrases in their e-mail signature block (“C’est la vie”).

 

SUMMARY

Foreign language documents will add time and cost to an eDiscovery project lifecycle.  You can mitigate these side effects by marshaling the right people, process and technology.

For More Information: 

Daryl Shetterly recently recorded a 15-minute segment on “Multilingual eDiscovery Trends” with one of my partners, Bill Belt.  The discussion was moderated by Kristin Hansen at Lionbridge.  To view the presentation click here.

Learn more about e-Discovery and its impact for German companies, click here: E-Discovery.

by Daryl Shetterly || The e-Discovery Myth