In an increasingly electronic world, many companies are faced with the problem of how to review electronically stored information for litigation, and many of them are turning to predictive coding electronic discovery using text analytics. Promising more efficient e-discovery gathering at significantly lower cost, predictive coding analytics may well be the next big thing in electronic discovery. However, before firms turn toward predictive coding analytics, understanding when using such software is warranted is extremely important as well as when to use an e-discovery company.
Predictive coding analytics is designed to work with large volumes of documents that are too costly to review manually; firms may find that the software is not as accurate with smaller volumes. Predictive coding analytics uses intelligent algorithms to determine what documents are relevant and which are not by “learning” from a subset of the total volume. For this reason, the larger the document volume, the more accurate the results are likely to be.
Software is “Trained” by Senior Attorneys
Unlike manual review where lower level attorneys or paralegals are responsible for “first pass” coding determinations, predictive coding analytics is programmed at the start of the review by a senior-level attorney. The attorney works with the system to “train” it in what to look for as relevant and non-relevant to the specific matter. This process was time consuming in the past, but technology has significantly sped it up along with making the software easy-to-use for even non-technical attorneys. This has resulted in dramatic time and cost savings in a very short time frame.
Know What to Expect
Firms who choose to use predictive coding analytics must have excellent communication with the vendor. Different predictive coding systems can serve different purposes; for example, some are used in place of manual review, some are used to enhance keyword-based review, others promote “accelerated” review, and some have the flexibility to improve workflow in multiple ways depending on the requirements of the review. Demonstration of the workflow and explanation of how the system can be used should be required of the vendor to be sure that the firm, client and vendor all have the same expectations.
There is no doubt that predictive coding analytics will be less costly than manual review simply because manpower is more expensive than computer power. However, the concept of predictive coding is still relatively new, making it difficult to predict pricing. Litigators may need to shop around to find a vendor offering the best predictive coding analytics and support for their needs at a reasonable price.
One of the most important factors to remember when considering the use of predictive coding analytics is that it’s best to work closely with opposing counsel in advance. If both sides can agree to the use of predictive coding analytics, both can benefit from the advantages, and the likelihood of disputes over the methodology will be lessened.
Predictive coding is one of the most discussed factors in electronic discovery today. Because it’s a relatively new technology, some firms are reluctant to be the “test case” to see if predictive coding analytics will be a viable alternative to manual review. Many firms believe that judges prefer the keyword search approach, despite the fact that U.S. Magistrate Judge Andrew Peck recently pointed out that keyword searches had never been sanctioned by the courts either. For more information on when to use predictive coding analytics, visit www.advanceddiscovery.com.