Before you file the support case, build the one-slide repro

When something breaks in a course, the fastest way forward I've found is to stop debugging the whole course and rebuild the problem on one slide.

 

A broken interaction in a 40-slide module could be the trigger, the variable, the layer order, the LMS, or the phase of the moon. The same interaction on a blank slide with nothing else on it has maybe three suspects. If it breaks there too, you've got a clean repro to hand support — and if it works there, you know the problem lives in the course, not the feature, and you start adding pieces back until it breaks again.

 

The part I keep relearning: the repro has to be minimal, not just small. One slide, one trigger set, no extra variables, no leftover layers from the real course. Every extra piece is another suspect you have to clear.

 

Curious how others do this — do you build the repro first, or only after you've already spent an hour in the full course? And what's the bug that finally taught you to isolate first?