A couple thoughts: You're zoomed out quite a bit on the horizontal timeline scale, and if you have a very short transition you're unlikely to get the cursor lined up close enough (and for long enough) to drop it on the clips. Zoom in, and also make sure you've got the clips actually touching each other. It looks like they might not be abutting each other in your screen recording.
If none of those work, take a look at this similar
discussion.