One is having good benchmarks prior to training. Most companies monitor production issues like inventory, scrap, rework, etc. A good ESD program would be performing frequent audits and recording deviations.
Hard data on performance improvements would show things like a drop in scrap, fewer boards in rework, fewer ESD control deviations, and so forth.
There are also more subjective measures - employee focus group feedback, employee satisfaction surveys. (There are strong correlations between availability of training and employee retention).
We also get a lot of feedback from team leaders and other managers. We often get reports that while on MBWA (management by walking around), employees are actually heard in discussion about concepts from the training courses.
Retesting after a couple of months to ensure concept retention is also a good idea in evaluating a training program.
But there is no substitute for the hard data. One large company recorded a 5% drop in scrap rate. All due to the training?? Probably not, but there would be grounds to suggest the training was at least partly responsible.