The increasing complexity of cyberattacks requires classification systems capable of representing how techniques interact and evolve. This paper introduces a relational matrix composed of eight attack groups and twenty directional connections that capture how one type of attack facilitates another. The matrix addresses limitations of static classifications by explicitly modeling interdependencies observed in documented incidents. Unlike hierarchical schemes, it enables bidirectional analysis of attack progressions and supports the incorporation of techniques without altering the overall structure. By identifying common escalation paths, the matrix enhances threat anticipation, improves strategic planning and supports early detection. Its structure offers a scalable and adaptable method for analysing adversarial behaviour across digital environments.