Gone are the days when brokers or buy-side professionals picked up the phone to make bids and close deals. Instead, they “chat” through their Bloomberg Terminals. The more than 320,000 traders, investors, and other market participants with a Bloomberg Terminal use Instant Bloomberg, an instant messaging tool commonly called “Bloomberg Chat,” to view securities, share information, and negotiate trades.

With more than 20 million instant messages and 20 million e-mails sent every day, Bloomberg Chat is a rapidly proliferating form of electronically stored information. Because chats typically contain sensitive information, including trading and price data, they are often highly relevant for litigation, compliance, and investigations. Moreover, Bloomberg chats must be preserved and archived for possible review by regulatory bodies such as FINRA or the SEC.

However, unlike e-mail and other electronically stored information, Bloomberg Chat is a complex data source with unusual metadata fields. For example, every time a user enters or exists a chatroom, Bloomberg typically logs these entrances, number of participants, timestamps and other metadata that can complicate processing and review. Furthermore, chat transcripts can be rife with non-relevant event logs and disclaimers; they often miss context due their casual and free-flowing conversational and jargon nature, making it difficult and time-consuming to isolate important information; and have excessive duplication, escalating the amount of rework reviewers must perform.

When lawyers and eDiscovery practitioners consider how best to handle Bloomberg Chat for legal matters, it can be beneficial to consult early on and be guided through the process, rather than potentially waste time mid-way through the review. Here are some considerations:

  • Vendors should be able to accurately capture metadata fields for filtering, targeted searching and analysis to avoid pitfalls later on (i.e., not have considered a metadata field unique to Bloomberg Chat). Many eDiscovery tools are ill-equipped to handle chat, and convert chat communications to e-mail file formats that lose chat-specific metadata fields, such as chat room ID and participant count.
  • Text file presentations of Bloomberg message logs are often easier and faster to review, since they are in a single “conversation” versus disparate documents. For example, a text file might read: “Phil enters the chat session. Phil says how are you today, Phil leaves the session.” In another format, it might read as three separate documents.
  • Make sure that attachments are being captured and not missed.
  • Because Bloomberg Chat includes highly informal language, jargon, lax spelling, and grammar shortcuts not often detected by traditional keyword search, linguists can consult upfront to develop a cost-effective and targeted culling, deduplication and search strategy to quickly eliminate non-relevant or duplicative data, and quickly hone in on relevant data.
  • Many eDiscovery platforms cannot support types of analyses needed for Bloomberg Chat, such as date gap analyses, thus necessitating custom data analytics approaches.
  • A standalone tool generally should not be required; Bloomberg data can be loaded into a review platform alongside other ESI for further search, review, analysis and production.

With a thoughtful approach to Bloomberg data, legal teams can surmount the challenges of Bloomberg Chat and reduce serious risks.