Apples to Oranges: How Competition Investigations Differ from a Litigation Approach When Implementing eDiscovery Solutions

By Daniel Rupprecht, Director – eDiscovery Consulting

An Overview of the Approach to Litigation Disclosure

At the outset of any dispute which is the subject of litigation there is a period of issue identification and information gathering.  During the course of this pre-trial procedure of evidence collection, commonly referred to as discovery, parties learn the facts that will provide courts with the information needed to reach a just and fair outcome for all those involved.  In modern day proceedings, much of the evidence needed to reach such a decision is located in electronically stored information (ESI). This means that technology is required to assist by digging deep into what is likely vast amounts of data from the various data sources that may contain material relevant to the matter.

Rules have been put into place to help guide this process, creating a balance as to what is reasonable in terms of the approach and cost, in order to satisfy the needs of both parties.  Typically, through negotiation, parties will outline relevant time periods and individuals of interest.  Data sources are identified and collected in an effort to create a universe of documentation where relevant information is most likely to be found.  Finally, once all is gathered, parties determine how best to interrogate said data and disclose the data that they are required to relating to the issues in dispute to the other side within a set time period.

Investigations: A Comparative Approach

The above all sounds like quite an orderly process –  but what happens if there is no clear indication of what the issues are?  How should parties proceed if there is no clear understanding of who might be involved? What should the approach be if a party is vast organisation with immeasurable amounts of data stored all over the world? When parties are conducting an internal audit, or are subject to a regulatory request from a government authority, this is the typical situation in which they find themselves, and demonstrates why such investigations differ so greatly from the litigation scenario outlined above.

Investigations require parties to think more creatively, as often no clear roadmap is easily identifiable when formulating an initial plan.  Investigations are not a one size fits all process where relevant information can be ascertained at the push of a button.  Parties want to get started and find the information needed as swiftly as possible, but the reality is that often no one knows where to begin – in data terms, that is not a good place to be.  Standard review models from the litigation approach, such as search term application, are often ineffectual because the quantities of data returned could be so vast as to preclude any value.  Legal teams need technology to assist them to identify the starting line, and to work with service providers who understand the nuances of investigation to ensure that time and cost are not wasted on inefficient strategies.

The good news is there is no need to reinvent eDiscovery, or learn a whole new set of skills to help drive new workflows.  The knowledge, tools and approaches for litigation matters can be used creatively to create a review set more in line with what would be expected in a more typical litigation discovery case.

Strategic Applications of eDiscovery Technologies

Several years ago I was a member of a team of lawyers hired to represent a client who had recently been dawn raided.  The client had no understanding of any nefarious behaviour that their employees may be engaged in, but as authorities are not in the business of knocking on doors without due cause, there was a level of concern.  Ultimately half a million documents across a ten-year time frame were seized and the corporation wanted to identify whether or not they should file for leniency.  The organisation was raided by the competition regulator in the jurisdiction, however, knowing the type of matter was not enough to determine the best options to start an investigation.  We needed to identify relevant information and lines of enquiry quickly.

In addressing the “when” in any given scenario, it is important to consider methods to narrow the scope.  Unlike in litigation, at this stage the goal is not to find everything, it is to find something – A starting point.  Knowing this was a competition investigation, the legal team requested the dates of various trade shows over the course of a year.  The reason being that events such as trade shows provide a forum and opportunity for anticompetitive behaviour and are thus a good starting point to get attorney eyes on documents quickly.  We isolated communications two weeks before and two weeks after the specific dates creating a tiny subset of documents to review first.  This strategy ultimately led not only to an understanding of what the infringement under investigation might have been, but also who was likely involved.

A Further Example Scenario and Solution

On another matter in which I was involved, our legal team represented an end client who was being investigated for cartel behaviour relating to market allocation.  Our team was provided with data from over 150 potential custodians from all over Europe.  The volumes presented were insurmountable in terms of time and cost to review in its totality.

As with understanding the “when” above, when determining “who” was involved, creative measures were needed to find a starting point.  Market allocation usually involves splitting up regions and creating agreements to compete, or (more likely) not compete.

A technology known as domain parsing was applied to all the data.  We were able to very quickly identify who had been speaking most frequently with competitors and created a small subset of documents to review first.

Again, this strategy proved invaluable. It not only provided clarity on who was involved and enabled us to shrink the initial custodian pool significantly, but also provided great insight as to how individuals within the market spoke to each other, and the relevant terms and vocabulary in use – enabling the investigating team to create more accurate search terms to identify further information of relevance.


The above examples are just two strategies that have proven to be effective.  They are not solutions that would work for every case, but given the right situation, could be used strategically to help isolate key documents quickly and create a roadmap where once there was too little information to get started.

Conceptual clustering, creating exemplar documents for categorisation, and predictive coding can all be used to help guide data interrogation when there is little information to start an investigation.  Even more basic analytics such as deduplication and email threading can significantly reduce volumes while at the same time they can also be reverse engineered to find out who knew what and when.

Having spent over 10 year working on some of the largest investigations in the world I have never failed to emphasise the importance that all members of a team, both internal and external, be completely aligned.  When working in the law firm, I found that the approach was far more successful (and significantly less frustrating) when partnering with external support who understood the requirement for an investigative approach, as opposed to partners who simply applied or focussed on their typical eDiscovery for litigation model. Trying to compare the two approached is truly like trying to compare apples to oranges.