The process of designing a pedagogic model for work-based learning within a virtual university is not a simple matter of using ‘off the shelf’ good practice. Instead, it can be characterised as an evolutionary process that reflects the backgrounds, skills and experiences of the project partners. Within the context of a large-scale project that was building a virtual university for work-based learners, an ambitious goal was set: to base the development of learning materials on a pedagogic model that would be adopted across the project. However, the reality proved to be far more complex than simply putting together an appropriate model from existing research evidence. Instead, the project progressed through a series of redevelopments, each of which was pre-empted by the involvement of a different team from within the project consortium. The pedagogic models that evolved as part of the project will be outlined, and the reasons for rejecting each will be given. They moved from a simple model, relying on core computer-based materials (assessed by multiple choice questions with optional work-based learning), to a more sophisticated model that integrated different forms of learning. The challenges that were addressed included making learning flexible and suitable for work-based learning, the coherence of accreditation pathways, the appropriate use of the opportunities provided by online learning and the learning curves and training needs of the different project teams. Although some of these issues were project-specific (being influenced by the needs of the learners, the aims of the project and the partners involved), the evolutionary process described in this case study illustrates that there can be a steep learning curve for the different collaborating groups within the project team. Whilst this example focuses on work-based learning, the process and the lessons may equally be applicable to a range of learning scenarios.