You cannot improve what you do not measure. But it is not always easy to tell what is worth measuring and what only adds complexity. Here is what years of practice programs have taught us about the metrics that matter.
Some metrics appear in every training program without being suitable for steering it. Others are easy to take; others, like the qualitative measurement of self-awareness, are hard to collect and tricky to analyze, especially at scale. After several years of refinement, we consider eight metrics, or observation angles, as the ones that define a balanced, efficient and effective program on critical conversations. They split evenly between results and discipline, because you cannot understand why results happen, or do not, without knowing how people actually trained.
| Result metrics | Discipline metrics |
|---|---|
| Quality of the strategy | Reliability |
| Self-awareness | Focus |
| Self-efficacy | Effort |
| Learning agility | Balance |
Skillgym measures more than 50 detailed behaviors, aggregable on any competency model, and asks trainees along the way how the program is going for them. But these eight are the 20% that carries 80% of the impact.
1. Quality of the strategy
In leadership conversations, confidence is the degree of courage, determination and perceived experience with which a person approaches a conversation, especially a critical one. If I had to pick one metric to predict the outcome of a conversation, it would be this.
We measure confidence as the overall quality with which the skills required by a conversation are used in what the trainee says, in context: the specific situation and the way the conversation develops. Defining it this way also defines the scope of the training: increasing people's confidence in taking and handling more leadership conversations.
2. Self-awareness
Everyone agrees it matters; the question is how to measure it. We ask trainees to rate their own performance and compare the rating with the evaluation produced by the analysis of the conversation. The closer the two, the higher the self-awareness. We do it in two steps: first the trainee's perception of their own behavior (internal self-awareness), then their perception of how the other person experienced it (external self-awareness). Taking both measures during and after each conversation builds a picture of development over time.
3. Self-efficacy
In a conversation, self-efficacy is the ability to keep a consistent approach and quality of relationship from start to finish, whatever happens. It affects the long-term quality of relationships in a team, and it affects the other metrics: a lapse in self-efficacy lowers the quality of the strategy and clouds self-awareness.
We measure it through what the trainee says: whether their approach stays coherent when the character pushes back, or swings abruptly. Self-efficacy should always be read together with quality of the strategy, because it is possible to be perfectly steady while running a bad conversation.
4. Learning agility
Learning agility is finding yourself in a new situation, not knowing what to do, and figuring it out. We find three dimensions useful. Skill flexibility: how trainees adapt a skill to a changing context, since what works with a direct report may not work with a peer. Learning speed: how fast they understand, apply and evolve their behaviors. Retention: how long a newly acquired approach lasts without maintenance. Aggregated, they say a lot about the efficiency of any learning strategy.
These four result metrics are tightly connected. You cannot work on the quality of the strategy without working on self-awareness, and so on. Reading them together, as the program runs, turns intangibles into measurable assets. But that is only half the story.
5. Reliability
You cannot expect results if people do not show up. Reliability is the ratio between the expected schedule and the sessions actually attended. It matters most in digital learning, where skipping costs nothing. We have found that turning sessions into calendar appointments, with reminders, is the single most effective lever on this metric.
6. Focus
Showing up is the beginning; then you have to be present. Focus is relatively easy to measure, through elements such as engagement during the conversation and the time spent reviewing a past one. In our data it correlates strongly with the achievement of learning goals.
7. Effort
Effort is the amount of practice, the easiest metric and one of the most used. Time and again, quantity of practice relates to mastery, provided the other metrics are in balance. Change does not happen for free. But consistency and rhythm matter more than intensity. It works like a gym: three short sessions a week beat one heroic afternoon a month.
8. Balance
A well-balanced program scores higher in results and engagement than one where quantity and type of practice are left undefined. Balancing means deciding which conversations to practice, how often, how many repetitions, when, and what happens in between. The best way to keep a program balanced for each person is an adaptive protocol that sets rhythm, sequence and difficulty from the trainee's actual attendance and results. The more personalized the path, the easier consistent engagement becomes.
Focus on what generates impact
Time and again we see a direct correlation between managing these eight metrics and obtaining faster, easier results. You can still track the classic competencies of a leadership curriculum, and we do: every behavior we measure maps on the client's own model. But your internal clients, the leaders you want to develop, have a shrinking attention span and little appetite for a competency map. Metrics they perceive as actionable are not only good L&D practice; they make the training more appealing.
Two notes on who sees what
First, every metric above is produced from what trainees say, by voice or by text, and from their own self-assessments. Nothing else enters the measurement.
Second, the detailed picture belongs to the learner. What the organization needs to steer a program is the aggregate view: reliability, effort and balance across the population, and the trend of the result metrics over time. That is what a program dashboard should show, and it is enough to understand which factor is moving the program as a whole.
Going digital makes all of this possible: some of these metrics can be collected by hand, others are too laborious, and a few are impossible without a system doing it for you. And it is not about measuring once. It is about measuring consistently, so that you can adjust the recipe while the program is still running.