In October 2024, IDC’s FutureScape: Worldwide Future of Industry Ecosystems 2025 Predictions forecast that by 2026, 50% of marketplaces would expand community capabilities, using fair-value monitoring to incent collaboration, and grow their data and application inventory. The other 50% would see participation decline. Two years later, Jeff Hojlo, Research Vice President at IDC, says the split roughly holds.
The route there looked different from the one the prediction described, and he now expects the next couple of years to land closer to 80-20.
Why did IDC call it a 50-50 split?
The call rested on how fast ecosystems were widening. Organizations have always worked with a small, finite set of partners, but in recent years the breadth and scope of those partnerships grew sharply. IDC’s 2023 Data as a Service and Data Marketplace Survey found that more than 80% of respondents considered external data critical to the success of every enterprise.
Hojlo ties that appetite for outside data to several pressures: ESG goals and the energy transition, the need to innovate faster with fewer people and thinner skills, customer experience optimization, and the move to software development. As he puts it, “everything is becoming software defined, whether it’s a product, whether it’s a building, whether it’s a process.”
Marketplaces already supplied data and applications, along with the APIs to connect them. Using them as a place to work with partners was a smaller habit. IDC’s 2024 Future of Industry Ecosystems Global Survey found that just 22% of respondents considered the marketplace a place to work directly with their partners and customers. In Hojlo’s account, the trend lines and IDC’s conversations with end users pointed toward 50%. Wanting external data and collaborating inside a marketplace to get it are separate habits, which kept the call from going higher.
The prediction also named what would divide the two halves. According to the report, organizations often take part in several marketplaces at once, and IDC expected participation to fall wherever fair-value monitoring was missing. The report called it one of the top roadblocks to a successful ecosystem.
What played out differently than expected?
Hojlo says usage moved the way IDC expected: roughly half of organizations now use marketplaces to engage partners and customers more closely and to co-innovate, as well as to pull data and applications. Who delivered the community features is where the story changes.
Major infrastructure and platform providers have not broadly added community management and co-innovation features such as sandboxes. They expanded AI model gardens and data marketplaces for sharing and marketing applications, and they added ways to make sure participants are paid for their work. That bears on the fair-value piece of the prediction. Fair-value monitoring has not emerged as such, in Hojlo’s view: no one is checking whether revenue is allocated fairly to partners on a given project. Providers do have strong incentives to draw participants and bring them back.
Enterprise app and industry-specific providers picked up more of the collaborative features. Their customers and partners need to share data, co-innovate on products and services, and source data relevant to their own domain. Hojlo has seen this most in the industrial space, where participants want targeted contextual data to complement what they already hold.
AI sits underneath much of this growth. Hojlo describes marketplaces as the supplier of the living data model that feeds the language models behind AI tools, and organizations now look to them for external agents as well as data and applications. Beyond agents embedded in applications, companies want outside sources that support their use cases and help them respond to disruption.
Does 50-50 still hold?
For today, yes. Looking a couple of years out, Hojlo sees closer to 80-20. The 20% would be marketplaces without payment mechanisms, a fair-value sharing approach, or a rich, contextual set of data and applications. In his words, “I expect all marketplaces, if they want to survive, will have some level of that.”
That outlook changes what “participation decline” means. Hojlo now frames it as a stickiness question: how well a marketplace gets people to stay and come back. There is room for both kinds of marketplace, in his view. A community-oriented one is built to keep partners returning to co-innovate. A transactional one, where someone grabs a data model, an application, or a software development resource, is a quick visit by design.
What should marketplace participants watch next?
That difference in purpose shapes where marketplaces compete next. Hojlo says differentiation still comes from data sharing and from community and co-innovation environments, whether a marketplace serves one industry or runs as a transactional model garden. Both depend on the data sharing infrastructure underneath. A newer requirement is the ability to look across multiple data sets and to orchestrate agents and data models with governance, so participants can find the best next step when an opportunity or an unexpected disruption appears.
The obstacles to getting there are mostly operational. In an IDC data as a service survey from last year, the challenges he describes were predominantly operational, not technical:
- Privacy and compliance concerns, which act as the first gate on whether data can be shared at all
- Infrastructure cost
- Token management and allocation across partners and participants
- IP management and legal complexity
- Confirming that a partner is who they say they are, through standard cybersecurity tools plus a validation mechanism and legal framework
Delivery matters just as much. On top of open APIs for connecting new data sources and models, database integrations, and data clean rooms to cleanse and operationalize data, Hojlo points to a workflow-ready approach and to monetization and incentive structures for the work partners do together.
The takeaway: decide what your marketplace is for
Hojlo notes that marketplaces have operated for hundreds of years, going back to medieval France. A successful one, physical or digital, offers breadth of options and a dynamic mix of partners serving different purposes, plus the architecture to support co-innovation and ongoing learning. For anyone building or joining one, his advice starts with privacy and compliance, which decide first whether data can be shared at all. From there, settle which need the marketplace serves and build data sharing and governance to match, with incentives that bring partners back.
