Contributed by MarkLogic
By Ed Downs, Head of Global Solutions Marketing, Financial Services at MarkLogic

Fast-paced, ever-changing market conditions are driving the need for better ways to do investment research. The pressure for better research comes from multiple angles. Actively managed funds are under increasing pressure to outperform the market, while  regulations such as MiFID II (Markets in Financial Instruments Directive II) require research to be unbundled and differentiated.

Meanwhile, digital wealth management platforms are winning younger customers with investment services at lower entry points. All of this is happening in a market that is more and more automated, with the human element more removed from decisions and predictions run by machine learning and AI.

One of the unavoidable truths about the future of investment research is that it needs to be faster. Data practices have fallen behind, and investment researchers must now find better technology and methodologies to keep up with demands from both the market and consumers.

MarkLogic Survey Data Results

68% of survey respondents reported that their research information is stored in too many locations. Only 34% reported being able to access all data in one place.

Finding Your Data
Many of today’s financial firms have data scattered across multiple data silos. It is usually kept in a data lake or other source that makes it difficult to pull together and analyze. In a recent survey by WatersTechnology[1], 68% of respondents reported that their research information was stored in too many locations. This makes it difficult to make good investment decisions, which negatively impacts performance.

In fact, poorly organized research data is a major impediment to effective research. When data is spread across disparate sources, it becomes hard to find and easy to exclude or pass over. This leads to poor conclusions and lots of guesswork. Previous generations of financial analysts could write this off as working on pure instinct or experience, but modern analysts need to justify their investment decisions in order to meet regulatory and compliance requirements.

The Promise of Machine Learning
In a market increasingly driven by machine learning, disconnected data has also become a problem. Today analysts are dealing with investment complexities that dwarf market landscapes of the past. It’s nearly impossible to digest and process the vast amount of market information available today without the help of machine learning. But those machines aren’t going to learn data that they can’t easily access. Incomplete input leads to faulty conclusions. Financial analysts need to know not only what data is available, but where it’s located. When data is scattered across too many locations, it’s too common and too easy for important research to be ignored.

The use of machine learning is on the rise, but with poor quality data, results are about what you would expect. Although 57% of the WatersTechnology survey respondents reported using machine learning, only 16% reported that machine learning has produced “promising” results.

Can We Fix Financial Data?
Trying to make sense of the sheer volume of financial data is daunting. It’s constantly streaming in from multiple feeds – whether it’s from the market or from internal organization sources – and there’s so much of it. At one point, data lakes seemed promising, but they have not surfaced as a viable solution. Financial analysts today are collecting more data than ever before, and data lakes are not making this data more accessible or actionable. However, there are ways to make sure that the data is both accessible and actionable.

Using a Data Hub to Increase the Value of Investment Research Data
A harmonized view of data is critical to providing valuable insights in the modern financial market. One promising approach is a data hub, running on top of a multi-model NoSQL database. This method of storing and accessing data differs from traditional approaches such as relational databases or data lakes. Some of the key reasons for adopting a data hub include:

  • Multiple data silos are consolidated in a manner that makes information homogeneous and easily accessible. As such, data hubs are often used to first optimize existing investments in data warehouses and lakes.
  • The lineage of individual records, providing easy access to every change in addition to the original source of incoming data, is maintained. This means that decisions can be based on solid, provable facts instead of hunches, guesswork, and old-fashioned instinct.
  • New methods of recording and analyzing connections between records, with semantics and knowledge graphs, is now possible. Without the strict structures of a relational database, a data hub is able to connect information in new and innovative ways.

The WatersTechnology survey respondents reported that none of their organizations were using data hubs for financial market research. 56% reported that such projects were in the works, which will lead to interesting results in the near future.

Instant access to the highest quality data is foundational for financial firms to develop an information edge and improve overall performance. Given the clear advantages delivered by data hubs, the question financial firms ought to be asking themselves is not whether they can afford to embark on such an initiative, but whether they can afford not to.


About MarkLogic
Data integration is one of the most complex IT challenges, and our mission is to simplify it. The MarkLogic Data Hub is a highly differentiated data platform that eliminates friction at every step of the data integration process, enabling organizations to achieve a 360º view faster than ever. By simplifying data integration, MarkLogic helps organizations gain agility, lower IT costs, and safely share their data.

Organizations around the world trust MarkLogic to handle their mission-critical data, including 6 of the top 10 banks, 5 of the largest global pharmaceutical companies, 6 of the top 10 publishers, 9 of the 15 major U.S. government agencies, and many more. Headquartered in Silicon Valley, MarkLogic has offices throughout the U.S., Europe, Asia, and Australia.

For additional information: