Transforming soft skills into observable behaviors

A competency model is only as good as the thing it is anchored to. Ours rests on two. The structural anchor is Gary Yukl's functional taxonomy of leadership behaviour, which sorts what a leader actually does into four domains: task-oriented, relations-oriented, change-oriented and external. The four-domain structure is Yukl's. The fourth domain we call system-oriented and scope towards accountability within the wider organisation rather than outward representation alone - both the name and that scoping are ours, not his, and it is worth separating the two. A functional taxonomy was a deliberate choice. It describes behaviour in a situation, which is something you can point to in a transcript, rather than personality, which is not. The second anchor is the healthcare context. The competency descriptions are informed by the World Health Organisation's competency framework for health workers, so that values-driven leadership, patient safety and systemic accountability are part of the vocabulary rather than an afterthought. Informed by is the accurate phrase: we use the framework as a reference point. We do not claim WHO certification, endorsement or compliance of any kind.

Adaptability as a single score tells a clinician nothing they can act on. Broken into four dimensions and observed across seven distinct situations - a high-emotion one-to-one, an experienced colleague refusing an instruction, a corrective conversation about burnout, a sudden reassignment, a priority reset with a peer, feedback inside a team, a sensitive treatment conversation with a patient - it becomes a map of where adaptability actually fails for that person. Adaptability is the one competency we describe here in full, as a worked example. The complete matrix of all 18 stays internal, and the limitations section further down says plainly why.

During calibration we split what began as a single empathy construct into two separate competencies: Empathy and Active Listening. It sounds like hair-splitting until you read the reports. One is social intuition - registering what the other person is feeling and responding to it. The other is communicative precision - actually catching, checking and reflecting back what was said. People are routinely strong in one and weak in the other, and a merged construct hides exactly that gap. The split cost us something: two competencies to define, calibrate and assess instead of one. What it buys is feedback that points at something specific enough to practise.

Every scenario is tagged with the function of the conversation before anybody is assessed in it. The reason is straightforward: assertiveness in a conversation about task precision does not look like assertiveness in a conversation about a damaged working relationship, and marking them identically produces feedback that is technically consistent and practically useless. Which competencies carry the most weight in which function is part of the internal scoring configuration. What matters publicly is the principle: the same behaviour is read against the goal of the conversation it appeared in.