Calculating the ROI of eDiscovery tools
By Paula Kulig, AdvocateDaily.com Contributor
Lawyers who have been slow to embrace eDiscovery technology tend to focus on the upfront cost while ignoring the considerable savings that come as the process progresses, says Jason Bell-Masterson, director at the Toronto branch of legal technology company Epiq.
“These technologies allow you to save money. You have to make an investment upfront but, as in most things in eDiscovery, you will often see significant savings in time, effort and money later on down the road,” he tells AdvocateDaily.com.
In addition to the upfront cost, he says, it’s difficult for Epiq and other companies that provide eDiscovery technology to be able to quantify what the ultimate savings will be, which acts as a further deterrent for some law firm clients.
“Most studies have shown that the majority of law firm costs — more than 70 per cent on average — come from the manual review of documents. Because you have to have someone put eyes on every document, linear review is labor intensive,” Bell-Masterson says.
“Therefore, the best return on investment from the use of technology, usually comes from the review phase and trying to reduce the cost of that aspect of eDiscovery. One way to do that is to speed up the pace of the review.”
There are a number of technology solutions that do that, but he says the most relevant are email threading, near-duplicate detection and technology-assisted review.
By threading group emails together by conversation, including the back-and-forth discussions between parties, it all stays together, Bell-Masterson explains.
“If you don’t use threading, those emails will be scattered among your entire document population. If you thread them, you keep them all together during review, and one person will see all of them at once, increasing both review speed and consistency.”
Near-duplicate detection looks at the text of standalone documents and “identifies groups of documents where the text is substantially similar — usually 80 per cent similar or more. What it typically pulls together is drafts. You might encounter them separately, but it’s often useful to see how the document progressed from one state to the next,” Bell-Masterson says.
“Email threading and near-duplicate detection are good ways to speed up the pace of the review. By grouping similar documents or email threads together, the reviewer looking at them can spend less time on each item,” he says, adding that there is now little resistance by law firms to pay for threading.
But bigger savings can be found using technology-assisted reviews, such as continuous active learning (CAL), Bell-Masterson says, which allows firms to more easily distinguish between documents that are relevant and those that aren’t —before reviewing each and every document.
“CAL and other predictive coding tools allow one person to review a statistically significant number of documents — usually a couple thousand documents. Then the software takes that one person’s review and scores the entire population of documents and assigns them a rank or number. You can then, at an early point in the overall process, say that anything in the entire document population above a certain number, such as 80 per cent, is likely to be relevant, and review those documents first, or even review only those documents.”
While the savings using technology-assisted review are substantial, he says, the cost is incurred upfront — unlike strictly manual linear reviews, where the costs are spread over the life of the project, sometimes for many months.
“The exact cost depends on the size of the case. The larger the matter, the more documents you’re dealing with and the more savings you will ultimately see from using these technologies. But because the cost is based strictly on the number of documents, and you’re paying it all at the very beginning, that’s where you can see the resistance.”
Bell-Masterson says eDiscovery technology, especially CAL, provides a major benefit for law firms.
“When you run this over your entire set of documents, the price is a fraction of what it would cost for a human review of those documents,” he says. “What you will typically see is as much as a 60 to 70 per cent reduction in the number of documents that would need to be reviewed. It’s absolutely massive.
"And the cost of running the technology is nowhere near what you would have to pay to review that additional 60 to 70 per cent of documents.”