Panel Discussion: Capture, Correlate, Analyse

Shield FC - Capture, Correlate, Analyse

Panel discussion hosted by The Realization Group, led by Shield Financial Compliance, Refinitiv, KPMG, Saxo Capital Markets & Lysis Financial, on 7th November 2018.


Firms are building libraries of data to be used for regulatory reporting. If built well, these repositories can be sliced and diced to give a variety of insights into the business itself, its activity and services.

However, those perspectives are only accessible if the underlying data is put together in such a way that it can be read by the wider business. That means it must be usable in context other than regulatory reporting.

On Wednesday 7 November 2018, at a breakfast briefing hosted at Refinitiv’s offices in Canary Wharf, a group of industry experts from across the financial services spectrum came together on a panel to discuss how to optimise this process.



It all begins at the point of capture. The capacity to reconstitute and understand trading-related data is dependent upon this initial stage, when the information must be cleaned, tested and validated.

Yet there is a tension. Within banks and asset management firms, senior managers are torn between implementation of cost reduction programmes – with data management typically being a significant cost – and the need to put a compliance process in place. Finding room for progressive investment in that process is challenging.

It is the chief officers within the operations and technology disciplines who must champion these programmes as business strategies, if they are to balance out those competing priorities whilst seeking competitive advantage.


The big catch

The first task for the leadership team is to bring compliance and the other elements of the business together, in order to set the foundations for the project.

Speaking at the event, Shiran Weitzman, Co-Founder and CEO of Shield Financial Compliance, said, “There needs to be some higher perspective in controlling oversight on all of these projects. Set up the groundwork, capture the data once and analyse the data once.”

The regulatory steering committee at a firm, which delivers the project to support the collection of compliance data and its use for reporting, should involve teams from across the business in order to help spot areas of data overlap, and to ensure that best practices are followed to support movement of data between functions.

Making this type of integration work is largely about culture, says Tom Kennedy, Global Head of Analytics Services at Refinitiv.

“Culture is vital in building an environment in which everybody can engage and collaborate – to communicate as a team,” he said. “It’s about building a good team mindset.”

On the technical side, the decision of whether to capture the data via siloed systems first and then move it to a single, central repository or to access the data directly from the siloed systems in real time depends on a firm’s use case.

Paul Burleton, Associate Director at Lysis Financial, said, “There are many issues with data being transferred from one place to another; you can lose information or get incorrect information transfer. That’s a massive problem, so try and get it from source, if possible.”

To prevent bottlenecks, and to facilitate the process of change, the underlying technology must be developed as an ecosystem of component parts rather than as big, monolithic systems that are joined by bridges. Using modular, microservice architectures can ensure that the ecosystem is flexible enough to handle the changes demanded of reporting, and yet can work with the wider business.

Causing correlation

Once the data is captured, it may still reside within different siloes. While that is enough to support reporting, to make it more versatile it must typically be drawn into other data analysis systems and correlated with other data sets in order to build a picture. This may not only mean rows and columns of numbers; it can also include unstructured data such as voice, the most challenging type of data to handle in this process, which must now be caught on the trading desk.

“This can be where artificial intelligence comes into play, when you need to understand the proximity of accuracy to know [the data] is indeed the right asset and we are correlating to the right order,” said Weitzman.

Mark Hollings, Associate Director at KPMG added, “You can ‘turn up’ your AI to be able to look for repeated patterns over which a Compliance or Risk Officer might zone out when surveying large data sets.”

Correlating interactions on phone and email with trading activity to support compliance processes such as market abuse or ‘know your customer’ rules cannot be achieved without taking a holistic approach, which must be conveyed to all participants.

Rachel Przybylski, Head of Market Structure at Saxo Capital Markets, said, “The challenge I see is in bringing together the data science teams looking at the client side and having them understand the challenges we might be facing in the regulatory implementation. One will facilitate the other. It’s just about getting everybody on the same page.”

It is not only teams that must comprehend; a fundamental aspect is the ability for one system to accept data from another system. The idea of a utility solution within an enterprise was mooted at the event, as a way of significantly reducing cost. The counterpoint to this was the challenge of executing a project which needs consent from across multiple functions.

Hollings said, “How to create optimal client outcomes, rather than just looking at this as a compliance problem, is a key element to build out of the business case.”

Deep analysis

Analysing the data, once it has been captured and successfully brought together, creates the potential for a firm to better serve the customer, assist the regulator, and better inform business decisions.

Burleton noted that the advantage data has is the capacity to model outcomes, which takes the relationship into possible future scenarios as well, to identify opportunities and prevent problems.

Kennedy said, “There are lots of new techniques using simulation, such as agent-based modelling, which allows you to look at things like contagion risk, counter-party credit risk, model risk, in different ways.”

Analytics can provide insights that senior management can use as control and oversight, from customer experience to minimising risk and increasing surveillance capabilities. Given the enormous potential these tools can provide, when embarking on the journey the company must start from the right place.

“You do have to look at what the biggest bang for the buck is,” said Przybylski. “There’s so much out there that you can do, you must consider your clients. Where is the most value? Then spend your money wisely.”

Weitzman echoed that point, noting, “It’s very important for any firm to look into the challenge as strategically as possible and have a vision. It can start with a modest vision, but if you have a vison you can bring the platform and the vendors in to make it a reality.”


For more information on Shield Financial Compliance (a client of The Realization Group), visit