The emergence of natural language analytics interfaces driven by generative artificial intelligence (GenAI) models has accelerated enterprise initiatives to enable data democratization—making data available to business decision-makers without the need to train them to use business intelligence (BI) tools. It has also heightened the need for agreed semantic models and business metrics, as well as technologies that facilitate the sharing and consumption of data as a product. As I previously discussed, data as a product is the process of applying product thinking to data initiatives to ensure the outcome—the data product—is designed to be shared and reused for multiple use cases across the business as it enables enterprises to streamline and accelerate the delivery of analytics and artificial intelligence (AI) initiatives. The market for software that enables the design and delivery of data products is evolving rapidly, especially among providers of data catalog-based data intelligence software.
I assert that by 2027, more than 3 in 5 enterprises will adopt technologies to facilitate the delivery of data as a product as they adapt their cultural and organizational approaches to
domain-based data ownership. The market for software that enables data products is nascent, however. The ISG 2024 Buyers Guide for Data Products highlighted that while many software providers offer functionality that enables users to publish and discover data products, fewer address capabilities to monitor data product usage, evaluate performance metrics or develop and manage data contracts. For example, while 95% of providers assessed graded A- or above for a dedicated interface for the discovery and consumption of data products and 79% graded A- or above for a dedicated interface for the development of data products, only 42% graded A- or above for functionality to address the definition and identification of licensing options for data products, and only 32% graded A- or above for functionality to monitor the availability of data products. That said, many software providers assessed in the most recent Data Products Buyers Guide had functionality in the public or private preview stage of development, so I anticipate that these percentages will be very different when we conduct our annual update of the ISG Buyers Guide for Data Products.
This is in part because many data catalog software providers are in the process of adapting their products to better reflect domain-based data ownership and support the development
of data products by business units. Many enterprises have become acquainted with the concept of data as a product as one of four key principles of data mesh, a cultural and organizational approach to distributed data processing that also includes domain-oriented ownership, self-serve data infrastructure and federated governance. Domain-oriented ownership makes business departments responsible for managing the data generated by owned applications and making it available to other business units using standard and interoperable interfaces. More than one-quarter (27%) of participants in ISG’s Data and AI Program Study have distributed responsibility for data and insights reporting to business units.
A data product can be a domain-specific data set—the equivalent of what has traditionally been thought of as a data mart—but it can also be an algorithm, an (AI) or machine learning (ML) model, or a custom-built operational application. Regardless of the format, the defining characteristic of a data product is the application of product thinking in the development process to ensure that the outcome is designed to be delivered as a reusable asset that can be discovered and consumed by others on a self-service basis. Product thinking also requires data owners to provide instructions and service-level commitments so data consumers can feel confident that the data product is up-to-date and of sufficient quality to be relied on for business decision-making. This is fulfilled through the development of data contracts, which are created alongside the data product and form the basis of an agreement between the data owner and the data consumer about the nature of the data product and its intended use.
Data contracts should include a description of the data product, defining the structure, format and meaning of the data, as well as licensing terms and usage recommendations. A data contract should also define data quality and service-level key performance indicators and commitments. This is another area of work in progress for many software providers, with 21% of those assessed in our most recent Data Products Buyers Guide grading A- or above for data contract templates. The metrics generated by data observability form a critical component of the development and sharing of data products, providing the information by which data consumers can gauge if a data product meets their requirements in terms of a variety of attributes, including validity, uniqueness, timeliness, consistency, completeness and accuracy. However, only 16% of providers assessed graded A- or above for the definition and identification of data product service-level or key performance indicators.
Given the complexity involved in developing, accessing and managing data products, data product platforms should also be assessed in terms of support for AI to enhance and automate data product development, data product classification, data product consumption and data product management. This is also an area in which many capabilities are in the private or public preview stage. Only 32% of software providers assessed graded A- or above for the use of AI to enhance data product classification, while 16% graded A- or above for the use of AI to enhance data product development and 11% for the use of AI to enhance data product consumption. I recommend that all enterprises evaluate platforms that enable the development and delivery of data products with a view to streamlining and accelerating the delivery of analytics and data initiatives and improving trust in data used to make business decisions. Would-be adopters should pay careful attention to the status and relative maturity of available products, however.
Regards,
Matt Aslett
Fill out the form to continue reading.