Why now is the time to deploy a data marketplace – insights from Gartner

Data marketplace adoption is growing rapidly, with over two-thirds of companies interviewed by Gartner having deployed one already, or expecting to in the next 6-12 months. By centralizing information and securely sharing reliable and understandable data in the right formats through self-service for humans and AI, data marketplaces finally turn data into business value.

Gartner has already described data marketplaces as “essential” for delivering AI-ready data, and recently published a new Market Overview on the sector. This lists 24 vendors, including Huwise, who are active in the space, and provides insights to help organizations make the best choice for their specific requirements. In this article I’ve drilled-down into some of the key points in the report, as well as highlighting the critical trends driving the need for data marketplace adoption.

Data and analytics (D&A) leaders have always understood the importance of increasing data consumption across the business. Empowering employees with self-service access to the right data delivers better decision-making, greater productivity and increased efficiency. It lowers the workload on data and IT teams as they don’t have to spend their time generating reports for the business, freeing up their time for higher-value activities.

The advent of AI accelerated this need for high-quality, reliable and accessible data. Training Large Language Models (LLMs) on inaccurate or incomplete data leads to poor outcomes and risks AI success. Many organizations learnt this the hard way. AI pilots delivered promising results based on a small subset of data, but attempts to put them into production floundered due to poor data foundations.

Agentic AI is driving another step change in data access requirements. Given their autonomous nature, AI agents are designed to operate independently across end-to-end processes, without close human supervision. They need reliable and above all contextualized data in order to deliver on their transformative potential.

All of this makes effective, universal access to data a business imperative. Data marketplaces – centralized, easy-to-use spaces where humans and AI can securely search, discover and consume data provide the answer. Sitting on top of the existing data stack, and connecting to all an organization’s data tools, they provide a seamless interface for turning data, particularly data products, into business value.

Recognizing this, Gartner has covered the market extensively as it has matured. Building on this, it recently published its first Market Overview for Data Marketplace Solutions, surveying the options available to data leaders across the category. Recognizing our technology strengths and over 350 global deployments, Huwise is one of the 24 representative vendors listed in the market overview.

For me three things really stood out in the report – how agentic AI is driving deployments, the different types of data marketplaces now available, and a concrete definition of the capabilities that solutions should include.

Data marketplaces may have begun as places for humans to find data, but AI agents are now becoming their biggest group of users. In fact, by 2030 Gartner sees data marketplaces as being primarily implemented for agentic AI. That’s not to say human business users won’t be an important audience to cater for, but that the rise of AI agents will only scale if they have access to relevant, reliable and AI-ready data.

Driven by this need, a third (33%) of respondents to a recent Gartner survey said they’d implemented a data marketplace, with a further third (32%) expecting to adopt one within 6-12 months. 18% are looking to deploy, but beyond 12 months, meaning just 16% don’t have one in their plans.

Agentic AI brings new challenges to data sharing, particularly around context. Essentially, AI agents can now reason remarkably well. However, they can still hallucinate and give factually incorrect yet plausible answers when they don’t understand the wider question and the environment around it. This isn’t a question of feeding AI more data, but putting that
data into context. Humans, particularly experienced employees, have built up their understanding of the context around specific data and processes. AI has not. AI is only as powerful as the context it is given.

Data marketplaces are able to help provide this context, based on how data is being used by humans. Take the example of finance. If everyone in the finance team accesses a particular data product when producing quarter end reports, that shows AI agents that this should be the source they turn to when creating similar reports. And because data marketplaces understand what specific job roles are accessing data, they can provide detailed, tailored context to help AI agents make the right choice.

The growing need for data marketplaces is driving a lot of vendor interest in the category. Not only is there an increasing number of stand-alone, pure-play vendors (such as Huwise), but existing data management tools companies are getting involved. Companies providing solutions in areas such as metadata management, data governance and data cataloging are embedding data marketplace-style features to their existing tools.

However the strength and maturity of these integrated capabilities can vary, particularly around measuring usage or deploying AI to aid data consumption. As they are part of technical tools they provide an interface designed to be used by data professionals, not business users. More importantly, they are tied to an existing, single data management solution, risking vendor lock-in. In practice, most companies have a diverse data stack containing tools from a range of vendors, leading to potential lock-in and interoperability issues.

Depending on your environment, Gartner is clear on which category to choose. The report states:

“D&A leaders with multiple data management solutions (such as multiple cloud databases, multiple data lakehouses, separate metadata, data integration, and data quality tools) should select a stand-alone data marketplace solution for ease of implementation, integration and interoperability across ecosystems.

Choose an embedded data marketplace solution when the data sharing and collaboration scope is limited to one data management solution, such as an existing data lakehouse. Ensure that the marketplace capability requirements are fully met by your existing provider. “

Like many terms, “data marketplace” is used to cover a range of solutions. To help clear up the confusion Gartner provides its own definition, along with a list of mandatory features that companies should look for.

In terms of capabilities, it’s important to understand that a data marketplace provides the usage layer for your data stack. So it has to be able to bring together data from across the business, and make it easily available to all in ways that are understandable and trustworthy. 

For us, we see an internal data marketplace as equivalent to an e-commerce marketplace. It has to provide an intuitive experience to the users buying (consuming) data products, while also making it easy for sellers (data product owners) to list and market their offerings. Just as on an e-commerce site, administrators have to be able to manage security and access, understand what is selling, where people are dropping out of the sales funnel and constantly improve the experience to increase engagement from both data consumers and data providers. Without usage, any investments in data marketplaces will simply not deliver ROI – irrespective of the quality of the data within them.

This vision is reflected in Gartner’s list of mandatory features:

• Register: The ability for data producers to list their data products on the marketplace
• Search: Seamless search to ensure users can discover the best products for their needs
• Governed access: Granular access management in line with corporate policies to ensure security and to prevent unauthorized access 
• Share: Data products can be shared in a range of formats, with a variety of applications and through different protocols to maximize usage
• Usage tracking: Full monitoring and metrics on which data products are being used, where, and by whom 
• Change management: Support to constantly improve data products with full change management and lineage features

At Huwise we’re committed to constantly improving our solution to meet the needs of clients and their users (human and AI). Over the last 15 years we’ve gained extensive experience into what drives successful data sharing, and use this to help customers get up and running quickly and effectively, whatever their requirements. Initial deployments can take just weeks, accelerating time to value.

This focus on usage and ROI is recognized in a new Forrester Consulting Total Economic Impact™ (TEI) study that we commissioned. Based on analysis of four Huwise customers, the TEI found a 474% ROI, payback in six months and $6.5 million in net benefits over three years for a 10,000 employee company. Click here for the full TEI analysis.

Data marketplaces are central to increasing data sharing and consumption, unlocking value and underpinning agentic AI. Without the access to data they provide, AI agent deployments will struggle to scale, holding back their promised business benefits. As the Gartner report shows, there has never been more choice when it comes to data marketplace solutions – but also that the stakes have never been higher when it comes to selecting the right partner for your needs.