The need for buy-in from financial professionals at the grassroots level is one of the greatest challenges for data technology, according to multiple industry leaders who cited issues such as power loss, data accessibility and corporate culture.

Although the expansion of big data and artificial intelligence capabilities – and budgets – is accelerating, especially due to the pandemic, financial institutions continue to struggle to push for widespread adoption internally. 

«The human factor is probably a very underrated aspect of developing a data-driven organization,» said John Wu, associate director for business development, North Asia, Australia and New Zealand at clearing and settlement services provider DTCC at a virtual panel during the ASIFMA Tech & Ops Week.

«It’s probably easy to bring in a steering committee or chief data officer to streamline communication of goals. But this we feel will come to nothing if there isn’t any buy-in at the local or grassroots level.»

Data Silos

One of the major factors cited is the continued existence of data silos which has been sustained in part by concerns about the potential loss of control by local or regional units based outside of headquarters.

«This is even more significant where a sell-side firm has several data offices in different regions, which is usually a result of addressing data strategy within a particular market segment,» Wu said in the panel which was focused on financial firms in capital markets. 

«Working towards a target operating model for a data-driven organization is like piecing together a jigsaw puzzle. You can probably see what it looks like at the end but all the pieces that are required to be in place before that is difficult to visualize.» 

Data Democratization

And the free flow of data needs to be made two-way in order to effectively eliminate such silos. According to Bank of Singapore’s chief data and innovation officer Celine Le Cotonnec, «the only answer to data silos is data democratization», calling data policies that are focused on minimal sharing and a «need-to-know basis» outdate.

«Data democratization means that anyone in the organization can have access to the right data, and is empowered with the right skill sets and right tools to get access to insights that would help him or her make the right business decision on the spot,» Le Cotonnec explained, stressing that data ops need to be housed under the business side for this very reason. 

Data Quality

Once channels enable data internally to flow freely – and safely – the quantity in any large organization is immense. But quality continues to vary, even for those that are already unraveling silos and rebuilding linkages. 

«[Data is] digitized. But the quality is poor,» said Goldman Sachs' APAC head of core engineering data platforms and co-head of core engineering Jia Mao who recalled a joke at the bank that it had more than double the number of datasets – 90,000 – compared with employees. 

«Just having huge amounts of data is not really going to give you the competitive advantage that you need. What is difficult to give timely access of high quality data in a safe manner [and] the challenge is big as volume, speed, complexity, demand accelerate dramatically.» 

Behavioral Change

But even with all the pieces in place, employees will ultimately be the key participant for returns to be felt be it usage, sharing or storage of data. This requires behavioral shifts at the bottom-up level that cannot be thoroughly achieved by top-down pressure alone.

Cultural changes cited by panelists include the adoption of a common language on subjects like data models, instilling habits on data hygiene practices and the view of data as more of a «first class asset»

Learning by Doing

Both Bank of Singapore and Goldman Sachs believe that an effective catalyst for this behavioral change is a platform-driven strategy.  

Both banks have internally developed platforms with effective and user-friendly tools to empower employees to directly leverage data themselves for purposes like the creation of data dashboards.

«Yes, asking the right question is important but we shall not forget that analytics is an iterative process,» Le Cotonnec said. «It’s only by doing that you will be able to learn how to ask the right question. It is not by writing a word document with whatever business requirements.»

finews.asia is an official media partner of the ASIFMA Tech & Ops Week.