Chief U.S. Magistrate Judge Paul W. Grimm, Dist. Ct. Md., has suggested a low-tech way to lower the costs and burden of electronic discovery: Adopt the checklist used by prominent litigator Stephen D. Susman, of Susman Godfrey. The checklist is a list of suggested pretrial agreements that Susman proposes to opposing counsel. In light of the incredible cost of pre-production review of e-discovery, one of the most significant may be the "5 x 5" proposal:
"Each side must initially produce electronically stored information from the files of 5 custodians selected by the other side during an agreed period of time. Only documents which have a lawyer's name on them can be withheld from production and only if they are in fact privileged. Production does not waive any privilege and documents can be snapped back whenever the producing party recognizes they are privileged. After analyzing the initial production, each side can request electronic files from 5 other custodians. Beyond that, good cause must be demonstrated."
Reviewing the records of five custodians at a time would certainly be far less costly than employing small armies of associates or contract attorneys to review ESI in response to sweeping document requests. Grimm quotes U.S. District Judge David Campell, Dist. Ct. Ariz., who has used Susman's checklist in his court, on discovery excesses:
There is a disconnect between what happens in discovery and what happens at trial. Of the 5% of civil cases that actually go to trial, I would guess that 95% of the discovery is never seen or heard by the jury. A few examples: lawyers often mark hundreds of trial exhibits, and yet rarely are more than a few dozen used in front of the jury; depositions are sometimes read or played, but they are rarely dispositive evidence in the case (lawyers manage to get the key witnesses to trial), and even when used for impeachment, depos rarely affect the outcome of the case; interrogatory answers and requests for admission have never been used in 32 civil trials over which I've presided.
According to a survey of major cases (with litigation costs over $250,000) involving Fortune 200 companies conducted by the Searle Center on Law, Regulation and Economic Growth, only one-tenth of one percent of the pages produced in litigation are ever marked as exhibits. That metric taken with Judge Campbell's statement that only dozens out of hundreds of marked exhibits actually get presented to the jury, suggests that as little as one-hundredth of a percent of produced records ever get presented to a jury.
Judge Grimm quotes Judge Campbell on a possible cause of over-discovery:
Part of the problem is that most lawyers don't try cases. They cannot tell what will truly be relevant. They also are insecure--they feel compelled to turn over every stone for fear of missing some important fact. Experienced trial lawyers - especially criminal lawyers, in my experience--can see up front what will matter to the jury and they go get it.
The above discussion probably supports what most corporate counsel already believe: There's too much electronic discovery. But while Susman's checklist is a great idea if the other side will agree, what about situations where the other side won't agree?
One general suggestion would be to develop objective metrics on over-discovery that could be used at conferences and in articles, blogs and briefs to argue for more limits on discovery using proportionality or to call for further rule changes limiting discovery. More specifically, take closed cases that went all the way through trial and examine the documents that were actually admitted into evidence as trial exhibits versus the number that were produced during discovery. This would be repeating part of the Searle survey.
However, then proceed to examine the custodians who produced those exhibits and see how many of the exhibits could have been found in the records of the top ten custodians. To the extent that all or substantially all of the exhibits were found in the records of ten custodians, that would argue in favor of imposing a 5x5 rule even if the other side doesn't agree.
Individual anecdotes would be interesting but large numbers of data points could be more valuable in documenting overall patterns of excessive discovery.