subscribe to our newsletter sign up
Finding the needle in a haystack

Finding the needle in a haystack

Promoted by

How technology can assist with class action cases, which pose unique document review challenges for both plaintifs and defendant firms


AUSTRALIAN CLASS actions have attracted a lot of attention recently within the legal sector, with a number of firms trying to get a piece of the action. International firms are also monitoring the market eagerly and plotting their next move. Much of the focus has been on the basics of what class actions are, how they are funded, the types of claim being brought and future trends. But very little has been said about the best way to manage these types of actions using the technology available today.

Generally, these are complex actions involving multiple parties, with large classes that create document-heavy cases. Timeframes are often tight and the pressure to review the material is intense. This is where technology steps in to assist. Class action litigation poses unique document review challenges for both plaintiff and defendant firms. Defendant firms need to review large amounts of client material, much of which is irrelevant but collected during the typical process of casting a wide net.

Plaintiff firms need to quickly assimilate large tranches of documents received in discovery. Plaintiff firms also need to manage large numbers of clients and how they relate to the various documents.

In 2014,, in conjunction with a client law firm, undertook a case study to evaluate the effectiveness of manual reviewers compared with technology assisted review. In our study, the initial review of 35,000 documents, which had been undertaken by paralegals, had to be re reviewed by senior associates due to a large number of poor coding decisions. Our analysis showed the paralegals achieved a recall rate of about 50 per cent. In other words, half of the relevant documents were missed. Although it would be tempting to blame the junior reviewers, we find these quality results are typical across firms.

We then re-ran the 35,000 document review with one senior associate training the predictive coding system. By contrast, the senior associate, after reviewing only 4,500 documents using a technology assisted review, achieved a recall rate of 50 per cent. We were able to calculate that when the senior associate reviewed 10,000 documents, the recall rate would be increased to 80 per cent, which is a standard set out in judicial rulings.

Although having a lawyer look at every document is often set out as the gold standard, on-the-ground analysis shows that there is a significantly improved alternative at a lower price point. Both defendant and plaintiff firms need to show their clients and funders they can provide real value by executing cost-effective and efficient document reviews. Technology-assisted review is an important tool in the arsenal at a firm’s disposal to ensure they can meet this requirement.


Review platforms and technology assisted review offer the following key benefits:

  • The ability to review documents that are conceptually related as a distinct bundle. Studies show that when reviewing related documents together, reviewers can vastly increase their review speed and quickly become subject matter experts;

  • Algorithms that can locate key documents simply by entering desired paragraphs of text, without the traditional syntax match of keyword searching;

  • A system that learns based on the team’s coding decisions, and presents similar document sets to the review team as it improves its learned behaviour;

  • Near duplicate detection to help manage the constant duplicates and versions of documents that invariably creep into discovery review;

  • A customised thesaurus that provides insight into industry terms, project names and generally allows institutional knowledge to be quickly absorbed;

  • The capacity to manage large numbers of class participants seamlessly within the application.


e.Law is a legal consulting and legal technology service provider specialising in E-discovery. To find out more, please visit

Promoted content
Recommended by Spike Native Network