Either way, I enjoy them. I do, however, think the business model could be a lot more scalable. Content providers should seek to either use the same space for multiple games, or use the same game across multiple scenarios.
This would be a great use case for current-tech level VR I think. Plop people into a correctly scaled virtual escape room in and bam, you can provide and deploy new content at the rate you can design it (perhaps physical locations not even needed).
Alternatively, build a spaceship, or a dungeon, or something else that can be generalized to different levels in the same basic space.
To anyone in the boston area, or, I suppose, Sweden, check out Boda Borg for what I think is a more entertaining take on them involving physical challenges, really abstract puzzles, and rapid feedback of try-fail-repeat.