Buy-side firms are far from satisfied with their present data management strategies and are still struggling to deal with hoary old challenges such as data silos and legacy tech stacks that are inadequate for their modern needs, according to the latest polling from A-Team Group.
Only a fifth of respondents feel positively about how they are managing structured and unstructured data and none said they were very satisfied, the poll of viewers of the latest Data Management Insight webinar revealed. At the same time two-thirds said that fragmented infrastructures still posed the biggest headaches to their management of data and a little more than half cited older systems as an impediment.
The responses reveal a buy-side that is stuck in limbo as it seeks ways to navigate new business models and incorporate new technologies to keep them competitive. As the business models of the sector evolve and companies are forced to redraw their data management operations, buy-side firms are being forced to navigate a new set of challenges. These include the ingestion of greater volumes of data and the application of it to new use cases, such as the management of a broader range of asset classes in their portfolios.
‘Disparate Systems’
The insights emerged in polls asked during Data Management Insight’s “Best practices for buy-side data management across structured and unstructured data”.
Webinar panellist John Joseph, head of sales engineering at automation specialist Xceptor, said the results illustrated how the buy side was still operating “disparate systems” that are failing to manage a tide of data that is “not going to decrease”.
Generative artificial intelligence (GenAI) and modern data systems are likely to exacerbate these new challenges because they will increasingly bring new sources of data – much of which will be unstructured – into firms’ systems, added fellow panellist Gurprit Singh, global head of data and analytics at asset manager Partners Capital.
Singh said that GenAI offered huge opportunities to firms. Partners Capital, for instance, was combing over troves of largely unstructured data that the company had amassed over 20 years, but only 5 per cent of which had yet been utilised. He expects to generate insights from the task.
However, he warned that generated data is only 60 per cent accurate at best and that human validation and intervention is still required.
Greater Role
Christina Schack, ****head of data operations and strategy at Vontobel Asset Management, agreed. She said that unstructured data would play a far greater role in analytics, which itself would be in greater demand over time. She cited her own company’s experience of creating a framework that permitted decentralised management and storage but that has common policies, principles and processes that provide for consistency in the data.
Technology is a key determinant of how well firms can perform the difficult task of reconciling its structured and unstructured data and ageing infrastructure is the grit that will jam the wheels of any management system.
Legacy systems have enough difficulty with conventional modern data processing but they are unable to blend structured and unstructured formats, argued Sam Barber, head of data services product development at data management services provider Rimes. Companies like his are benefiting from this lack of on-prem capabilities because many buy-side firms can only achieve the necessary functionality by outsourcing many of their data processes. This helps them achieve what Barber calls “operational elasticity” – paying for processing only when it’s needed without having to invest in technology that might remain unused for most of the time.
More Obstacles
The panellists agreed that another obstacle that the buy-side faces in effectively using both structured and unstructured information is a lack of standardisation in data. This, said Joseph, is the common denominator between getting the most out of both types of data.
Many of the challenges presented to buy-side firms by the new data demands they face can be ameliorated with good planning at the management implementation stage, panellists agreed. Barber called for “abstract thinking” to ensure that all parts of the enterprise are considered when drawing a modern data management strategy, and Joseph recommended taking steady steps towards scaling any new framework.
Singh said his experience of working within the buy-side for the first time at Partners Capital convinced him that organisations would do well to take guidance from observations of how their peers had performed in their own transformations.
Schack closed the discussion by stressing the usefulness of AI in bringing harmony to structured and unstructured datasets. Without it, she said, companies are forced to use manual processes to bring structure to unstructured information – processes she said are still prominent. AI will also help in the vital task of improving data quality, she added.
- A recording of the full webinar is available to watch here. Visit the A-Team Group website for news of more Data Management Insight and other events.
Subscribe to our newsletter
