Training Metrics · Reference
Training metrics: definitions, formulas, and benchmarks
A reference catalog of the metrics used to measure training — what each one means, how to calculate it, and the benchmark to aim for. Grouped from the easy activity numbers to the outcome numbers that prove a program worked.
Training metrics fall into six groups: delivery and cost, then the five evaluation levels — reaction, learning, behavior, results, and ROI. The first groups are cheap to collect and weak evidence; the later ones take a baseline and a follow-up and carry the real proof. This page defines each metric. For the step-by-step method of capturing them, see the companion guide on training effectiveness; for the model behind the levels, see the Kirkpatrick model. Benchmarks below are typical targets and vary by program type and delivery mode.
Delivery & costOperational metrics
The share of enrolled learners who finished the program. An activity metric — it shows reach, not effect.
The share of sessions attended or activity logged, usually drawn from the LMS.
Total program cost spread across participants — development, delivery, time, materials, and technology.
Average time learners take to finish, useful for spotting overlong or abandoned content.
Level 1Reaction metrics
The average rating learners give the experience. Useful mainly when paired with the open-ended comment that explains the score.
Willingness to recommend the program, a single comparable number across cohorts.
Whether learners found the content relevant and which skills they intend to use first — a better predictor of effect than satisfaction.
Level 2Learning metrics
The increase from a pre-training assessment to an identical post-training one. The most direct measure of learning.
The share of learners meeting a defined competency threshold on the assessment.
How much of the gain remains at 30, 60, and 90 days — whether the learning stuck.
Level 3Behavior metrics
The share of learners using the new skill at work within 30 to 60 days. The most underused metric, and the strongest single predictor of effectiveness.
Sustained changed practice, confirmed by manager observation, self-report with examples, or 360-degree feedback against a baseline.
How quickly trained employees reach full productivity compared with a baseline or untrained peers.
Level 4Results metrics
The measurable gain in the business metric the training targeted — quality, sales, customer satisfaction, or output — attributable to the program.
The drop in mistakes, rework, safety incidents, or compliance failures after training.
The difference in retention between trained and untrained groups, often the clearest results signal for onboarding and leadership programs.
Level 5ROI metrics
The monetized benefit of the program set against its full cost, using the Phillips formula. Best reserved for high-investment programs with quantifiable outcomes.
A simpler companion to ROI — how many dollars of benefit each dollar of cost returned.
Picking from this list is a matter of how far you will act. If you only report the delivery and reaction metrics, you are measuring activity. The behavior and results metrics are where effectiveness is proven — and each one depends on a baseline and a follow-up tied to the same learner, which is the method covered in the training effectiveness guide and supported by training evaluation software.
