Handling the growth of data and the shift to Data-as-a-Service

By Martijn Groot, VP Marketing and Strategy, Alveo

The diversity and volume of raw data on financial products, marekts and companies for banks and financial services to work with is expanding rapidly. This is partially driven by the requirement for firms to disclose more information on transactions, investments and portfolios but also due to a growing need for transparency when it comes to environmental, social and governance (ESG) aspects of financial services. In general, data is also being generated and collected in larger volumes via growing digitalisation simply as a business activity by-product.

This growth in volume and diversity of data provides opportunities for financial services firms to extract market intelligence, improve operations and provide a better basis for competitive differentiation. However, to be able to leverage this data, organisations need to avoid not seeing the wood for the trees and be able to harness these growing volumes.

Hurdles and opportunities

New data sets can be provided raw or packaged in different ways with access typically via APIs or sometimes delivered directly to cloud storage or into a cloud based data warehouse as opposed to traditional, file-based delivery. Analysis often starts early on directly on raw content. For example, Natural Language Processing provides a means for businesses to be able to gain content from text-based data such as news, earnings calls transcripts, prospectuses or other filings. Instrument or entity identifiers lists or keywords can focus this textual data analysis and enable effective extraction of the content needed. Data set integration from varying sources can then facilitate curation or quality-control of data.

In the past, drawn-out data preparation processes were typically driven by monthly or quarterly reporting cycles, leading to insights that were either inaccurate, dated or both. The combination of processing data over a long period of time and relying on data that is ultimately poor in quality to drive business decisions will be insufficient to enable financial services organisations to keep pace with fintechs and challenger banks. Collection, aggregation, curation and analysis is now a continuous function.

To ensure that the data management function provides value, its objectives must be clearly defined. For example, it can focus on providing regular data sets for BAU operations or be more focused on enabling self-service data collection and analysis by scientists. Its remit can include cost management and optimizing Data ROI as well. Clearly defining remit and targets is vital.

What data best suits an organisation will hinge on markets, clients and geographies the firm currently has working relationships with, and will ultimately lead to lists of interest specifying collection requirements. Linked to that will be requirements around metadata, such as turnaround time in SLAs, usage permissions and restrictions or quality metrics. Thanks to technological advancements, data preparation and curation is becoming a faster process, with data harvesting, data set combining and live insights becoming a near real-time process. Use cases and the questions that require answering will help to drive the curation and collection of data.

A data analyst with the right capabilities is able to complete all of these tasks and bring the big picture-view to data to help the C-suite make business decisions.

Altering processes

The last few years have witnessed a shift in how data management and analytics processes are performed by financial services firms, which is providing a plethora of benefits to data scientists, quants and analysts. Previously, the two disciplines have had their stark differences, with the data management process covering cross-referencing, data sourcing, cleansing and reconciliations. Analytics are usually a subsequent process, facilitated by desk-level tools and libraries, which are close to users and operating on a disparate dataset.

The separation between these two disciplines has been a problem for financial institutions, with time-to-access being affected and slower decision-making processes. Now, the influx of cloud native technologies is enabling a more integrated approach between management and analytics.

One such example is Apache Cassandra. This distributed, highly-scalable and open-source database facilitates the ability for financial time series data to be secured and managed. Apache Spark allows for big data processing as a unified data engine, and together the two are able to bring data and analytics together, powering the decision-making process.

Organisations need to remember that any data being applied to decision making must also be of high quality, or otherwise run the risk of analytics not providing value, with inaccurate intelligence driving the strategy of senior decision-makers. With analytics in place, how do organisations then ensure these insights reach the right decision-makers quickly?

Towards Data-as-a-Service as a foundation for value-added data management

With the data management function working across analytics, covering more data sources and providing real-time capabilities, self-service can be provided to staff across different departments and more informed decision-making can be facilitated, with easy access to the data required.

Data management solutions suppliers are now moving into the managed services space, rather than just software. In addition, this is growing to include a Data-as-a-Service (“DaaS”) model, which includes data checks on top of suppliers hosting and running data management infrastructure. Dashboards can provide visibility into data preparation, while data formatting capabilities allow for last-mile integration with business applications.

Client-side staff are then able to spend time working on value-added data analysis. Data-as-a-Service is able to work on any data set, plus third-party data sources processing and benchmark and curves data, including ESG. Clean and prepared data will provide value in performance management, risk management and compliance.

With prepared data sets, quants and data analysts are able to attain key metrics that help to drive the senior decision-making process. Data scientists will be analysing data across different asset classes to fit certain criteria, such as ESG, to incorporate into the investment decision-making process. Increasingly, AI and machine learning is also being incorporated, alongside innovative data science solutions.

By adopting this mindset, proprietary analytics, with an endless combination of data types can facilitate stress-tests, valuations, investment decisions, performance analysis and risk management. With the necessary detail and information then filtered to C-suite decision-makers, data scientists can be a vital cog in business strategy. Additionally, self-service capabilities, which allow the reviewing of lists or requesting of sources can reduce the data supply change cycle.

Ultimately, Data-as-a-Service solutions can provide a solid foundation to the operations and analytics of a business. Data quality is increased and the change cycle is shortened, benefitting business functions. When adding this to quality metrics on data sets and sources, continuous improvement can be applied to a data operation.