A Systematic Review and Taxonomy of VR Usability Heuristics: Trends, Challenges, and Directions

Abstract

As Virtual Reality (VR) becomes prevalent in high-stakes and everyday applications, ensuring seamless user experiences (UX) is essential. While Heuristic Evaluation (HE) is a highly valuable usability inspection method, its application across VR domains and devices remains fragmented and underexplored. This review examines VR usability research from 2004 to 2025, drawing on 53 rigorously selected studies. We propose a methodological taxonomy categorising heuristic approaches into four distinct types: VR-Specific and Custom Heuristics (41.5%), Combinations and Mixed-Methods (26.4%), Traditional/General Heuristics (24.5%), and Unspecified evaluations (7.6%). While Core HCI and Technology remain the foundational domain, HE has rapidly expanded into Education, Industrial Training, and Healthcare. Hardware usage predominantly involves immersive Wearables (HMDs), with hybrid 2D/3D systems emerging strategically in clinical and industrial contexts. Crucially, traditional 2D heuristics often fail to capture VR's unique spatial, physical, and cognitive complexities, including simulator sickness, multi-user latency, and spatial UI clutter. Recent literature (2024–2025) reveals a critical shift toward mixed-method evaluations, pairing HE with standardised metrics such as the System Usability Scale (SUS), and validated, domain-specific frameworks (e.g., for wearable VR educational apps). By addressing these gaps, adopting mixed methods, and refining frameworks, the community can deliver intuitive, safe, and user-centred VR experiences, unlocking its full transformative potential.