Closing The Loop

Your eDiscovery software should be able to find "e-discovery," right? But many discovery platforms cannot. If you search for “e-discovery” in most platforms, your search is likely to come up with inaccurate results—even if you’re searching documents full of the term.

Of course, this failing isn’t just for “e-discovery.” Many platforms don't register single-letter words or hyphenates like e-mail. They’re thrown off by terms such as “not.” Even searching for numbers can go wrong—all of which can greatly complicate your investigation.

So, why is software that’s designed to help discover information so bad at searching in these instances? The issue lies with the technology that these platforms are based on.

For example, if you’re attempting to run a keyword search in Relativity, one of the largest eDiscovery technologies, the platform will, by default, ignore a host of “stop words.” These are words and characters the search technology ignores “because they do not act as meaningful criteria in a query.” That shunned, ignored, forgotten text includes punctuation marks, single letters, single numerals, and a multitude of words—112 by default—ranging from “about” to “must” to “your.” Punctuation like periods, dashes, and colons are also ignored. Thus, “e-discovery” is treated like “e discovery,” the single letter e is ignored by the search technology and your keyword search for “e-discovery” goes horribly, ironically awry.

The same is true for advanced searches. Relativity, and platforms based on it, use the dtSearch engine to conduct fuzzy searches, proximity searches, and the like. Built in the early 90’s, dtSearch was an early text retrieval software used in eDiscovery, search engines, web crawlers, and more.

But it, too, can’t find “e-discovery” easily. The engine requires an alphabet file to tell it which characters to treat as text, spaces, word breaks, or to simply ignore. In the default mode, hyphens are removed when a search index is created. “E-discovery” again becomes two words, “e” and “discovery.” That’s not the only phrase that can be impacted by the search technology, either. Since punctuation inside a word is treated as a space, contractions like “can’t,” “would’ve,” “it’s,” are treated like two separate words: “can” and “t”, “would” and “ve”, “it” and “s”, etc. and ad nauseam. (Emphasis on the nauseam.)

This is why, as Craig Ball recently pointed out at Georgetown Law’s Advanced eDiscovery Institute, not just Relativity, but most eDiscovery platforms cannot find the word e-discovery.

Of course, there are workarounds to this. That’s why Relatively has a 114-page search guide to supplement its 139-page user guide. But these workarounds are hardly simple and straightforward. They often require technical customizations to already complex processes. This shouldn’t be too surprising for eDiscovery products that have a 13-step process for running a simple keyword search, and whose difficulty and complexity has given rise to a whole industry of eDiscovery vendors. But they’re certainly not the sort of simple, powerful solutions that are needed to make eDiscovery accessible to all legal professionals.

There are, thankfully, alternatives to eDiscovery platforms that don’t know how to find “e-discovery.” Logikcull’s FlexSearch, for example, is based on the idea that legal professionals shouldn’t need a doctorate in information science in order to conduct a document review, that users shouldn’t have to learn or unlearn complex syntaxes just to create an effective search, that discovery technology should be usable, accessible, and reliable.

When you search in Logikcull, stop words do not interfere with your search. All words are indexed and searchable, so you can find them easily. The same goes for special characters, whether @ signs or accents. And if your query is invalid, Logikcull triggers an alert, rather than allowing an unresolved search.

And because we believe that discovery should be available to everyone, Logikcull adapts to you, rather than forcing you to adapt to it. Logikcull’s FlexSearch can identify your specific search style, recognizing when you’re using search behaviors common to Westlaw, or dtSearch, or Lucene. Logikcull then applies those behaviors to your data in order to deliver the best result.

It makes searching simple, easy and accurate, instead of a burden. And yes, it can find “e-discovery.”

This post was authored by Casey C. Sullivan, who leads education and awareness efforts at Logikcull. You can reach him at or on Twitter at @caseycsull. 

New Call-to-action

Subscribe To Our Blog

New Call-to-action

Let us know what you thought about this post.

Put your comment below.