Satisfaction surveys are the measurement tool most organizations use to evaluate training because they are easy to administer and produce numbers that look like data. They measure one thing: how participants felt about the training experience. They do not measure whether participants learned anything, whether they applied what they learned, or whether the application improved performance. Organizations that make training investment decisions based on satisfaction data are making decisions based on the wrong information.
A More Complete Measurement Model
The measurement model I use has four stages. Reaction — how did participants respond to the training? This is what satisfaction surveys measure, and it has limited value. Learning — did participants acquire the knowledge and skills the training intended to develop? Measured through skill assessments at the end of the training. Behavior — did participants change how they work as a result of the training? Measured through observation and structured manager feedback thirty to sixty days after training. Results — did the behavioral change produce the business outcomes the training was designed to support? Measured through operational performance indicators sixty to ninety days after training.
Each successive stage is more difficult to measure and more valuable. Most organizations measure only the first. Organizations that measure through to the fourth stage consistently make better training investment decisions.
Building Measurement Into Training Design
Measurement does not happen after training design is complete — it is built into the design from the beginning. The learning objectives that define what the training will develop also define what the learning assessment will measure. The behavioral outcomes that define what participants are expected to do differently also define what the follow-up observation will look for. The operational results the training is intended to produce also define the metrics that will be tracked in the sixty to ninety days following deployment. When measurement is designed alongside content, it is feasible and cost-effective. When it is added as an afterthought, it is expensive and unreliable.