This was the primary weak point in the implementation. Expecting users to recognize that they can click on the static-looking fields on the card (or relying on them to use tab to move between fields) does not pass the grandma test. It'd be great if you detected when a field had been fully filled out and moved to the next one automatically.

My experience on an iPad. Typed first 4 numbers of my VISA. Realized I could type in the MM and YY, then realized I could enter the name. It was a very smooth process for me - and I didn't read any text above or below before trying.

I wonder if trying it on a touch device made it more natural or not.