01
Step one: choose the skills and levels
List the skills the team needs to deliver its work, keeping the list between eight and fifteen items. Group them if useful, such as technical, client-facing and coordination. Then define a simple scale with written descriptions, for example one for learning, two for working independently, three for able to teach others. Written definitions matter because they keep ratings consistent between people. Worked example: a five-person support team lists billing questions, escalation handling, product knowledge, written replies and onboarding new staff, then rates each on the three-point scale.
02
Step two: collect ratings and check them
Place people in rows and skills in columns, then ask each person to rate themselves. Have the manager review the same grid separately, and discuss any gap of more than one level. The conversation is the valuable part, because it surfaces hidden strengths and unspoken assumptions. In the support team example, one agent rates herself two on escalation handling while her manager sees three, and the discussion reveals she has handled difficult cases that were never recorded. Update the grid with the agreed rating and note a date.
03
Step three: read coverage and act on it
Scan each column and count how many people rate at level two or higher. A skill with only one capable person is a risk, and a skill with nobody at level three needs a teacher or outside training. Turn each gap into a specific action, such as pairing, a course or a hiring need. In the example, only one agent can onboard new staff, so the team pairs two colleagues with her and schedules a review in three months. Revisit the matrix whenever the team changes.