client need

A multinational technology company was experiencing frequent regulatory scrutiny– especially in the area of antitrust. With billions of dollars in fines and penalties at risk, the company’s compliance team began a regular series of anti-trust audits to unearth and correct potential risks. Each quarter the business was carefully checked using a method that included a range of specific and broad search terms and a review of more than 100,000 documents by contract attorneys overseen by the company’s outside counsel.

why Epiq?

It was necessary to search documents and emails for anti-trust issues, but with large random samples and an abundance of documents hitting on key search terms, the overall process was expensive and time consuming. The estimated cost was $150K per audit and $600K per year. The company’s compliance team needed new tools and workflow to improve the result, but lessen the time and expense required. This is where the Epiq technology-assisted review (TAR) team stepped in.

Epiq solution

Epiq revised the company’s regulatory audit to take advantage of advanced conceptual analytics, structured analytics, and predictive coding. We identified documents across the custodian collections that were conceptually similar to documents identified as high interest or relevant by earlier audits. Epiq also focused the contract review team on much smaller and more pertinent random samples. Rather than a very large random sample, the review was focused on a smaller, true random sample, a stratified sample (presenting a wider variety of types of content), and documents hitting on key terms. The real test, however, was in the effectiveness of the analytics tools in identifying relevant documents similar to those found in earlier audits.

results and benefits

In a side by side comparison, a review of 24,200 randomly selected documents by contract attorneys identified 20 actionable documents. The new analytics driven approach brought in a few random and stratified documents, but largely focused on documents conceptually similar to the previously found examples. Review of only 800 documents from the targeted set identified 44 actionable documents. In other words, the team found twice as many relevant documents but only reviewed 3% of the volume in the analytics supported approach.