By    |    October 20, 2020
Making compliance surveillance more intelligent and less costly begins with advanced logic.

Compliance officers are critical assets for options trading firms like PEAK6. From managing the extensive requirements of our industry to safeguarding us against risks like fraud, market manipulation and insider trading, our compliance officers sift through vast amounts of data every day. But spending too much time searching through imprecise piles of data—hunting for relevant information to review—was preventing them from prioritizing their focus on the issues that posed the greatest threats. We set out to change that, uncovering richer data to better identify risk, while saving our team time, and our company resources.
 

The old way of doing things

The standard compliance review process used to work like this:

  • Existing reports identified potential problems based on one or two data points, using simple logic.
  • The data was handed off to the compliance team to investigate.
  • Compliance officers would get each report; pull up the order, symbol or fill to examine it; and then our team of reviewers would determine if there was a legitimate justification for the flagged behavior.

This system of identifying problems meets regulatory standards, but it creates a huge time burden to identify what specifically in that data is relevant, concerning and erroneous. This is because the simple logic reports broad, imprecise potential errors.

Moreover, the overwhelming majority of investigations are cleared once the compliance officer demonstrates legitimate reasoning behind these flagged transactions. In 88 to 99% of these, one of roughly 10 commonly seen and easily quantifiable justifications could clear the flagged data. This was happening as a time-consuming manual process.

For example, an insider trading report might flag a symbol when a large move took place. Often we didn’t actually make any trades in that symbol leading up to it getting flagged. If we didn’t trade it, it clearly couldn’t have been insider trading. Yet our compliance officers were spending a significant amount of time that resulted in these types of common, quantifiable outcomes. Instead, we wanted to focus our time and analytical skills on the rare and harder-to-spot data that presented real risk.

A great leap forward

We introduced advanced logic. Instead of using one or two data points like in the past, the new reports now use about 65 data points.

  • The updated, advanced logic reporting provides exact data points and reasoning why it flagged something, including why that trade might be excused.
  • Each trade is assessed and cleared as appropriate based on these updated criteria.
  • If a trade or data point fails to meet these criteria, only then does it get flagged for manual review.

In short, when the data can be excused for a common, quantifiable reason, the new reporting automatically examines and clears situations in which this applies. Commonly flagged data no longer need to be excused manually. All of this reduces workload considerably, enabling compliance officers to use their time more strategically to spot emerging and high priority risks.

The new and continually improving way

Our advanced logic has reduced the number of manual reviews needed by as much as 99% per report. This doesn’t reduce the need for a compliance officer review, but it allows compliance officers to focus their time and analysis on flagged data that wasn’t cleared by the system and has more significant risk potential.

In addition, we’re continually looking to improve the system and add reports. We’re planning to introduce machine learning to identify data that may not fall within our strict guidelines, but could still be relevant.

  • We have created a separate set of logic that identifies items that don’t fall within our parameters for review, but are “close”.
  • Reviewers look into that data and determine whether or not it is relevant.
  • Once we compile enough data from these close transactions, we will train a machine to identify what is and isn’t relevant, without the strict guidelines of the advanced logic. Ultimately, this will enable us to review even more valuable information.

These improvements continue to better our compliance team’s ability to focus on the most relevant information at any given time, while reducing the time needed for report review and keeping our firm safer.


Want to learn more about our tech-enabled compliance team? Check out our open compliance roles.

Adrian Dalton

Adrian Dalton

Adrian Dalton is a compliance officer and technologist at PEAK6 where he develops in house RegTech. Adrian uses his experience and education in law and technology to create legal and compliance applications to improve efficiency through reliance on data, advanced logic and machine learning.