Squeezing the most value from data has become the key driver of data management innovation in the past few years. Among the tools garnering most attention in this quest is an approach that treats data as a consumer product.
The theory is a simple one. By packaging datasets as well and data-centric services and products, such as analytics models, as items to be bought by consumers other than those who developed them. the constituent data can be more efficiently distributed and used.
Cloud platforms such as Amazon’s AWS and Microsoft’s Azure already host digital outlets for such products. And vendors including S&P Global offer marketplace tools.
Data product, and the emergence of data marketplaces from which they can be accessed, was the focus of much debate at A-Team Group’s Data Management Summit London 2025. The concept was central to the keynote presentation given by Justin Arbuckle, Global Head of Enterprise Platforms and Practices, Schroders, which was entitled “Getting data products right”.Data products also figured in a keynote address given by State Street Bank’s Sunny Jaisinghani, Head of Data, Global Custody Product, and Aneet Shah, Senior Vice President, Global Head Custody Product Development, under the theme “Future proofing the data eco-system and enabling a front-to-back experience”.
And Tony Seale, Knowledge Graph Architect at The Knowledge Graph Guys, grounded his keynote “From data deluge to actionable knowledge: Leveraging data products & Knowledge Graphs” in the concept of the data product.
Definition Division
Summit speakers had differing views on the exact definition of a data product but broadly agreed that as well as being created for “reuse” they are characterised by their cross-organisational application, which in turn needs a stable and, where possible, standardised interface. Importantly they must be traceable and discoverable, a necessity that has given rise to data marketplaces.
Data products also serve a strategic role as catalysts of collaboration, fostering a culture of data sharing and reuse. And while it is likely that competing products jostle for space in data marketplaces, each is created to serve a specific purpose.
To be effective, the products must also be trusted and properly catalogued to ensure that they can be distributed and used as widely as possible. A data product that has no “buyer” is a waste of resources, one observer noted.
New Products
The summit heard a proliferation of products had hit marketplaces, largely comprising datasets based on customer insights, financial metrics and operational performance data. Among others were machine learning training sets, Internet of Things and sensor datasets as well as compliance and risk datasets.
Data products, however, are subject to an array of challenges that are commonly faced by all data – the inability to be optimised within fragmented legacy systems and their vulnerability in the face of inconsistent data. Creation and adoption of them is also prone to a unique set of other obstacles, a prominent one being organisations’ resistance to embracing data reusability and sharing frameworks. Overcoming such a radical cultural change in a sector that has hitherto jealously protected its data and IP is proving difficult for some firms, the Summit was told.
Looking Ahead
Data products can play an integral part in firms’ efforts to future-proof their data ecosystems. The Summit heard how flexible data governance strategies, agile tech infrastructure and adaptable corporate cultures are the cornerstones of a data setup that can weather changes in the trading and operating landscape.
Along with the need for high-quality data, these are all fundamental attributes of data products.
Standardised, consistent and reusable datasets available in a self-service framework can provide many of the building blocks that a “modular” data management system needs to adapt to evolving markets and economic conditions, the Summit heard.
Those same attributes will also benefit the construction of knowledge graphs, a concept that is among the latest approaches data representation and configuration. Interest in knowledge graphs has been fired by the search for the optimum way of arranging data so it can make best use of artificial intelligence.The Summit heard how knowledge graphs, by stressing the intersecting relationships between different data points, can enable more to be done with that data and provides for greater connectivity across siloed datasets. In that way, they provide for better data connectivity, which will develop more meaningful AI training.
While data products – by their standardised, clean and interoperable nature – help in the construction of knowledge graphs, they are in turn being improved by their inclusion in these network-based structures. They can be better connected throughout a data network and those held within fragmented systems can be more optimally accessed and utilised, the Summit heard.
Subscribe to our newsletter
