Business communications are evolving from paper to electronic format, thus complicating matters of privacy, tractability and organization. In order to help users make sense of and deal with these electronic communications during discovery, systems must handle the vagaries of our modern world.

Not so long ago, written business communications were relatively formal and structured. People wrote letters and memos, and created reports, all on paper. For informal communications, handwritten notes might be jotted down, with the expectation that they would be read and swiftly disposed of, leaving no trace.

The proliferation of email in offices changed all that. Suddenly, the line between formal and informal communication exchanges blurred – often a single email message might incorporate both formal and informal information.

There were three important consequences to the arrival of email.

First, concepts of privacy changed. Documents created on a business computer system belong to the business, not to the individual writer. Just as your teacher could force you to read aloud the note Sue just passed to you, so can the company, a judge, or an opposing party in a lawsuit read your email.

Second, electronic documents rarely exist in a single place. Copies end up on computers throughout the network, so disposing of them is not a simple undertaking.

Third, instant messages (IMs), along with text messages and emails from mobile devices, are messier – less structured and less formally written – than many other forms of communication. Abbreviations and "creative" spellings abound. Understanding any individual IM requires knowledge of the context of previous communications between the sender and recipient. Without this context, outside observers will find the dialog cryptic and vague.

Traditional techniques

Most software sold for electronic discovery has not come to grips with the modern world of communication. It is based on simplifying assumptions about communication exchanges. The most salient assumptions are:

• The context of any document can be understood just from its content, without any need for any external reference.
• Words are correctly spelled, without abbreviations.
• Language is clear and unambiguous, without any implicit reference to external context.
• Communication exchanges are linear. You can find the whole exchange by retrieving the email thread that carried it.
• Communication patterns are aligned with predictable, documented business processes and systems.
• Language is clearly understandable, without any jargon or specialist terminology.

In reality, these assumptions are flawed. Any technology that is based on them will have significant and potentially fatal shortcomings in terms of its ability to help attorneys with review or with developing and pursing case strategy. Any software that is adopted in the hope of addressing corporate needs in an age of changing communication technologies needs to address numerous real world issues.

Many communications do not make sense on their own, instead requiring the context of other communications to clarify their meaning. At the very least, seeing documents in context makes for faster and more accurate decisions about their importance and relevance.

Many communications, especially informal ones, contain misspellings or abbreviations.

By their very nature, email messages and IMs are often vague and rely on implied meaning. Sometimes, these communications are intentionally ambiguous, such as when the author is concerned about the implication of what he or she is writing.

Any email or IM dialog is likely to be informally structured, at best. Within a single email thread, topics may change such that the original subject is no longer being addressed. This may happen gradually as the topic drifts from one issue to another. Or a change of topic may be abrupt, when a recipient of a message simply clicks "reply" to write to the original author on an unrelated subject.

The converse situation can occur where a topic may be explored across numerous email threads among different people. A simple email thread tool would not be able to correlate these various conversations as being related. More sophisticated analysis of the content and context would be required to do this.

Sometimes, a conversation may be conducted using a variety of different communication vehicles. Participants might switch from email to instant messaging, or from a business email system to home email. Where records are available from these sources, it can be invaluable to be able to find related items that span multiple systems.

Above and beyond the contents of the communications, reviewers and investigators may gain valuable insight through analysis of communication patterns. Knowing who communicated with whom can be an important indicator about whose communications should be subject to particular scrutiny or, conversely, that certain individuals may not be pertinent to a given matter.

Or communication patterns may reveal, for example, that a key executive was not a party to conversations about an area for which he or she was nominally responsible. Such analyses can provide important insight into the true power and decision-making structures of an organization, which can often differ from formal organizational structures.

Every industry has technical terminology, abbreviations, and jargon that are not generally known outside that field. Furthermore, organizations – and even individual departments – develop yet more specialized vocabulary. To conduct accurate analysis, a system must have some means of incorporating such terms.

Systems need to be able to deal gracefully with communications that involve various forms of deliberate obfuscation, including circuitous language or even switching to foreign languages to communicate sensitive information. Choosing a system that can handle the unexpected interjection of a foreign language is one step in addressing such situations.

More sophisticated still are systems that can recognize the tone of a communication; for example, do the author's words indicate fear, anger, or evasiveness? Detection of such qualities when topics of interest are being discussed may highlight communications of particular significance.

Simple tools for finding information in written communications have shortcomings even for well-structured collections of documents. Most computer users have encountered keyword searches that return a multitude of irrelevant documents while missing crucial ones. These shortcomings are exacerbated when the data under scrutiny is informal and obscure (sometimes deliberately so).

In order to help users make sense of and deal with the real world of communication during discovery, systems have to be designed to handle the vagaries of the real world.