Many people, I think, work in the world of formal training for a time, then gradually broaden their perspective. Courses and classes, organized instruction seem to make sense for getting people started with basic skills and knowledge that they lack. When Amtrak designed a completely new reservation system, it planned training for the 2,000-plus employees who worked with the previous one.
I’ve heard “These people need training” at the start of countless projects, and often they did need some training. Usually they needed something else as well, and at times something other than training.
Joe Harless, a past president and longtime contributor to ISPI, loved grappling with on-the-job performance. Along with his enthusiasm and insight, he brought a deceptively quiet approach to helping a client explore an apparent problem.
“Never ask a client, ‘What do you want people to know?'” Joe advised.
Why not? “Because he’ll tell you. And usually what he wants people to know is something like appreciation of widgets, and the history of widgets, and great widget makers, and widget policy, and…”
Instead, Joe said, the question to ask is, “What do you want people to do?”
The difference between the two questions may seem obvious to you, but for me it was like being in the ocean after only having seen pictures.
Even if the client’s model of training involves only lectures and PowerPoint, “What do you want people to do?” shifts the focus to the reason they’re on the job — the results they’re supposed to accomplish. (If the client focuses only on how they perform, you can ask about the results they produce — whatever’s left over when the workers go home.)
The magic in this question is that by starting with the desired outcome, you can more easily explore with the client different factors affecting that outcome. When the client believes that “training” is a solution (and when you actually do have a skill/knowledge gap), the conversation can include how to maximize the impact of the training and how to address other barriers to performance.