The recent RIMES forum in Toronto, held at FIMA Canada, addressed many data challenges that buy-side firms are encountering. During the forum, moderators presented on the results of both the RIMES 2014 buy-side survey and also research conducted by Gartner.
Participants observed a presentation on these subjects, and got a chance to give their two cents on matters such as data lineage and data architecture. One major point of discussion was the different models for centralizing – or spreading out – data, along with their respective challenges.
RIMES 2014 buy-side survey
The presentation highlighted the results of RIMES buy-side survey 2014, which polled more than 100 participants across all major financial centers worldwide.
In addition to drawing replies from key decision makers in various regions, the poll garnered responses from end-users working in a wide range of job functions. Staff involved in data management, IT, performance, compliance, risk, front office and operations took part.
The survey detailed the varying priorities of central management staff and end-users. While 48 percent of the former identified centralization as a priority, only 26 percent of the latter had this point of view.
While end-users were less likely to be worried about centralization, they were more concerned about cost control – with 74 percent specifying this as a priority. Only 59 percent of people involved in centralized management highlighted cost control as a concern.
The RIMES forum moderator commented on the situation, asserting that end-users end up paying the cost either way, while central management may not be in the same boat.
Gartner research findings
After going over the results of the RIMES survey, the presentation reviewed the findings of the recent research conducted by Gartner, emphasizing that there are several variables that must be considered. For example, there are three dimensions, which include:
- Complex relationships: This involves the interplay between different markets, legal entities and financial instruments
- Velocity: This pertains to the speed of data
- User diversity: Not only business users, but also regulators, customers and unknown users harness the information contained by buy-side firms
The forum covered the different kinds of data architecture, as well the difficulties that come along with these approaches. Companies can spread their data out, but this can lead them to end up with different versions of the truth.
Alternatively, many companies centralize their data, but this approach can certainly generate challenges for the various end users at buy-side firms.
The presentation noted that a company’s architecture frequently hinges on the organization’s size. For example, many smaller firms rely heavily on outside data. The problem here is that the key information is outside of their control.
As they get larger, the data is still not within their control, but they are bringing in user layouts. As companies move toward having all their data be provided internally, they get closer to shared information.
RIMES summed up Gartner’s view, stating that when companies have shared information, users can demand a greater amount of applications and data. This can lead to latency, and users can become disenfranchised.
These employees might decide to do it themselves, and get information on their own. In addition, the data hub may become idle because nobody wants to use it.
Amid this situation, Gartner is frequently encountering a hybrid approach to data management.
One participant, a global information architect for a major financial services firm, described this as being a “federated” approach. The attendee said that in this situation, “That which needs to be centralized is centralized and governed, that which can be done by the states, would be done by the states, because of their responsiveness to local leads.”
He emphasized that there are always going to be challenges, no matter what approach you use.
“Everyone wants an easy solution, but there is no easy solution.”
Data lineage is another area where buy-side firms can frequently run into problems. For example, if users obtain their information from a single data source and then convert it, they may have a hard time getting the same end result repeatedly. A participant from a major financial services institution chimed in once again, noting the challenges his company has faced in this area.
He emphasized that while his company has many tools it can use for data lineage, they don’t work across all the different platforms and the different means an organization can use to Extract, Transform and Load.
“There [are] a variety of disparate approaches to the lineage at present,” the attendee said. “Every time we look for a solution, we are going to be data-centric, or we are going to be streamlining processes. I wish it was either or. Unfortunately, it’s both.”
“The maturity in this space is still growing … and the lineage is very definitely an issue because no single tool is going to solve that problem,” he stated.
While buy-side firms are encountering many challenges in the current environment, there are several tools they can use to overcome these difficulties. Since data architecture solutions are complex, institutions might benefit from getting an in-depth assessment from RIMES. By doing so, they can identify the ideal approach for their challenges.