According to Science Daily, 90 percent of the data in the world has been created in just the last two years. That is a sobering statistic for corporations trying to control the cost and risk of their litigation discovery. Amid this data explosion, however, there is good news. There are multiple ways to bring order to chaos. The question is, which path to take?

Consensus on what works best

There is some emerging consensus at a high level about how to deal effectively with the “electronic discovery” challenge. Most repeat litigants would agree that e-discovery nirvana involves setting up a repeatable, standardized set of processes from data identification and preservation through processing, review and analysis to production and presentation.

Atul Gawande, in his book The Checklist Manifesto, does an outstanding job of explaining why. He shows how effectively delivered checklists in emergency medicine have had a dramatic impact on results (i.e. fewer people die). If production of privilege or missed discovery deadlines are the litigation equivalent of medical disasters, then process, tied to the right talent and technology, is where to invest. Corporations that have put in place well-defined e-discovery processes have dramatically increased their litigation discovery effectiveness, reduced their risk and saved tens of millions of dollars into the bargain.

Ways to get there

Typically, there are three models that corporations adopt:

Preferred provider. In this model, corporations select a small number (two or three typically) of vetted preferred e-discovery providers. Rates are generally negotiated up front and are generally pay-as-you-go. Outside counsel is informed at the start of the matter which provider they will be working with. All the corporation’s discovery data is now in a select number of outside locations (as opposed to highly dispersed).

Sole source. In this model, a single provider is selected to handle a set of e-discovery functions, typically on a multi-year basis. A set of custom processes and reporting mechanisms are developed and constantly refined over time with the corporation client. Often, a dedicated team is put in place, either at the client site or at the service provider site. And in some instances, the corporation buys software that the service provider runs and manages. Payment models range from variable pay-as-you-go to fixed fee. All the corporation’s discovery data is in one single location.

Build. In this model, the corporation builds out its own e-discovery function. The corporation buys software and invests in the infrastructure, talent, training and process building required to deliver e-discovery services to internal and external counsel. Internal teams can be anywhere from five to 30 people.

Making the right choice for your company

As with many questions in life, there is no one right answer. Corporations that pick the “build” model tend to have cultures that demand absolute control or believe they have a core competency (or can build one) in the area of data management. The “preferred provider” model tends to appeal to corporations that want to bring some order to the chaos but are not ready to get married to a single provider. The “sole source” model fits for corporations who want one single set of processes, metrics, and people working their matters, without the up-front and ongoing investment required to build a team and manage technology internally.

Going back to Gawande’s powerful, data-driven stories, the “sole source” model, all other things being equal, should lead to the best outcome for most corporations. By definition there should be one optimal set of (constantly improving) processes. For example, once you’ve found the optimal way to screen for privilege or remove junk files prior to review, why have two, three or five of these processes? Similarly, corporations gain great efficiencies from having one team with which to communicate and build a close relationship—sometimes referred to as “one throat to choke.”