Ask someone to list the steps of a process they were trained on last month and most will manage it. Put them mid-process with incomplete information and a clock running, and the training evaporates. That gap — between knowing about something and deciding well inside it — is exactly what scenario-based learning exists to close.
A scenario is a decision with consequences
The format matters less than the anatomy. A working scenario has a situation the learner recognises, a decision point with genuinely plausible options, and feedback that shows consequences rather than announcing "incorrect." The wrong options are the design work: each distractor should be something a reasonable person actually does, so choosing well requires the exact judgement the job requires.
Let them feel the cost, safely
The most effective scenarios let a learner make the tempting mistake and watch it play out. Recognising a problem too late, choosing the comfortable option over the correct one, skipping the step that seemed optional — the consequence lands harder in a simulation than in a bullet point, and it lands without harming anyone. That felt experience is what transfers to the real moment.
Where teams go wrong
Two failure modes appear constantly. First, the quiz in costume: a recall question wearing a story ("Sarah needs to know which year the policy was introduced…"). Second, branching for its own sake — sprawling decision trees that take weeks to build and test, when three sharp decision points would teach more. Complexity should live in the judgement, not the plumbing.
Descriptions inform. Decisions train. If the job is made of judgement calls, the learning should be too.