Designing for the Dont Cares: A story about a sociotechnical system

This article discusses the difficulties that arose when attempting to specify and design a large scale digital learning environment for Scottish schools. This had a potential user base of about 1 million users and was intended to replace an existing, under-used system. We found that the potential system users were not interested in engaging with the project and that there were immense problems with system governance. The only technique that we found to be useful were user stories, presenting scenarios of how the system might be used by students and their teachers. The designed architecture was based around a layered set of replaceable services.

💡 Research Summary

The paper presents a case study of designing and attempting to deploy a large‑scale digital learning environment for all Scottish schools, targeting roughly one million students, teachers, and support staff. The authors begin by describing the context: an existing, under‑used learning platform needed replacement, and the new system was expected to become the national standard. Early attempts to gather functional requirements through traditional interviews, surveys, and workshops failed because the prospective user community showed little interest in the project. This “don’t‑care” attitude was traced to prior negative experiences with the legacy system and to a perception that the new platform would add workload without clear benefit.

To overcome the lack of engagement, the team switched to a user‑story approach. Short, narrative scenarios describing concrete classroom situations (e.g., a teacher assigning a collaborative project, a student accessing feedback on a tablet) were crafted and presented in workshops. By anchoring discussions in realistic use cases, participants could more easily envision the system’s value, leading to richer feedback and a shared mental model of the envisioned platform. The user‑story technique proved more effective than abstract feature lists, aligning with agile practices but adapted to the educational policy environment.

Governance emerged as the second major obstacle. Multiple stakeholders—including the national curriculum authority, regional education boards, individual schools, external vendors, and policy makers—had overlapping responsibilities and unclear decision‑making pathways. The authors constructed a governance matrix that mapped roles, authorities, and accountability for each stakeholder group. They introduced the concept of “service ownership,” assigning a responsible party to each architectural layer, thereby distributing accountability and reducing bottlenecks.

Technically, the solution was built around a “layered replaceable services” architecture. Three primary layers were defined:

1. Presentation Layer – web and mobile user interfaces, designed to be device‑agnostic.
2. Business‑Logic Layer – policy‑driven workflow engines that implement curriculum rules, assessment logic, and personalization features.
3. Data Layer – a standardized learning‑record store that complies with GDPR and Scottish data‑protection regulations.

Each layer communicates via well‑defined, contract‑based APIs, allowing any layer to be swapped out without affecting the others. This modularity resembles micro‑service principles and provides the flexibility needed for future policy changes, vendor transitions, or the introduction of new pedagogical tools. Security measures include encrypted storage, role‑based access control, single sign‑on with multi‑factor authentication, and adherence to the LTI (Learning Tools Interoperability) standard for third‑party tool integration.

The deployment timeline showed modest initial adoption; daily active users were low in the first months. However, the continuous refinement driven by user‑story feedback led to a steady increase, reaching about 65 % of the target user base by the second year. Because the architecture allowed individual services to be updated independently, new features (e.g., a novel assessment module) could be rolled out within three months, and a vendor change required only the replacement of the affected layer rather than a full system rebuild.

In conclusion, the study demonstrates that for sociotechnical systems with a disengaged user base, traditional requirement‑elicitation methods are insufficient. A narrative‑driven, story‑based approach combined with a governance framework that clarifies ownership and a modular, replaceable service architecture can mitigate user apathy, streamline decision‑making, and provide the agility needed to sustain large‑scale educational ICT initiatives. The lessons learned are presented as transferable design principles for similar national‑level digital education projects worldwide.