One useful way to understand how companies develop and evolve their offerings is to look at what they fundamentally organize around. In my experience, most organizations fall into one of three broad categories: project-centric, product-centric or platform-centric.
Project-centric companies organize everything as a project. Whether they’re developing a product, adapting it for a specific customer or rolling out a new capability, the dominant mental model is the same: a clearly defined start and end date, a fixed budget, a predefined scope and a temporary team that disbands once the work is “done.” This approach feels reassuring because it creates a sense of control and predictability. Unfortunately, it also reinforces a transactional mindset where learning stops when the project ends and long-term evolution becomes someone else’s problem.
Product-centric companies represent a step forward. Here, the primary unit of organization is the product rather than the project. Each product typically has its own dedicated team, roadmap, budget and commercial goals. Products evolve over time through successive releases and there’s usually a stronger sense of ownership and accountability than in project-centric organizations. However, product-centric structures often struggle once the number of products grows. Each product becomes its own silo, with limited coordination across teams, duplicated functionality and architectural divergence that accumulates quietly until it becomes painful.
Platform-centric companies take a different perspective. Instead of treating each product as an independent entity, they invest in a shared platform that provides common capabilities across multiple offerings. Product teams build on top of this platform, adding product-specific functionality while benefiting from shared infrastructure, data and architectural decisions. Over the last few years, however, I’ve observed an even more radical evolution of this idea: the move toward what I call a superset platform.
A superset platform assumes that the full superset of all functions and features across all current and future products is generated from a single, coherent platform. In this model, individual products are no longer built as largely separate artifacts. Instead, each product is effectively a configuration of the platform: one specific instance chosen from thousands or even millions of possible combinations. The “product” becomes a particular selection of features, parameters, behaviors and constraints rather than a separate codebase or architecture.
An important aspect of the superset platform idea is that it extends far beyond software. In many software-intensive systems, especially in domains such as automotive, industrial automation and consumer electronics, the platform includes mechanical, electronic and increasingly AI-based components as well. Different component types naturally evolve at different speeds: software may change weekly, electronics annually and mechanical components over several years. A superset platform explicitly acknowledges these differences while still treating all components as part of a single, evolving whole.
This shift has profound implications. First, it fundamentally changes how scope is owned. In a superset platform, scope is no longer owned by projects or even by individual teams. Instead, it’s owned at the platform level, with decisions made based on global optimization rather than local convenience. Breaking architectural changes aren’t driven by the needs of one product or one customer, but are evaluated from a holistic perspective that considers long-term value, backward compatibility and the impact across the entire product portfolio.
Second, the superset platform enables a much tighter coupling between continuous value delivery and architecture. When products are configurations of a shared platform, improvements in the platform immediately benefit a wide range of offerings. Conversely, learning from one product instance can inform improvements that propagate across the entire portfolio. This creates powerful economies of scale in learning, not just in development.
Third, the model challenges traditional funding and governance structures. If products are configurations rather than standalone entities, funding “projects” or even individual products becomes increasingly artificial. Investment decisions need to focus on platform capabilities and value streams instead of predefined deliverables. This is uncomfortable for many organizations because it requires letting go of familiar control mechanisms in favor of intent-based governance and quantitative value models.
The transition to a superset platform isn’t easy. It requires explicit architectural leadership, a willingness to rethink ownership and accountability, and the discipline to manage variability and backward compatibility over long time horizons. But for organizations that aim to deliver continuous value at scale, across many offerings and markets, it’s becoming increasingly difficult to see a viable alternative.
The move toward superset platforms isn’t primarily a technical optimization. It’s a strategic shift in how organizations think about products, architecture and value creation. Products cease to be isolated bets and instead become expressions of a shared, evolving capability base. In the context of RADICAL, the superset platform replaces project-based and team-based scope ownership with a globally optimized foundation for continuous learning, adaptation and value delivery. To end with Peter Senge: “The essence of systems thinking is seeing wholes. It’s a framework for seeing interrelationships rather than things, for seeing patterns of change rather than static snapshots.”
Want to read more like this? Sign up for my newsletter at jan@janbosch.com or follow me on janbosch.com/blog, LinkedIn (linkedin.com/in/janbosch) or X (@JanBosch).
