I'm unclear on how you're drawing the scope of donors vs "everyone". It's, for all practical purposes, impossible to make a judgment about the moral impact of an action for everyone. There are so many secondary effects of every action that instead you're choosing two points in that space of impact and saying coffee vs orphans, which is really conceding the whole point. Sam took money intended for one purpose and used it for another. The purpose was embodied in the will of the donors and no other people.
Take for instance a simple second order impact, someone gets a free coffee of the card and decides that while it was a good coffee, he or she should instead donate 10x that to feed orphans. If more than 10% of people using the card do this, then there's a 1:1 follow on donation effect to the orphans. That is essentially the idea of "pay it forward" as far as I understand it. Sam decided that betting on his fellow humans wasn't a good one (perhaps knowing his own moral failings) and decided to short circuit the process and make decisions for others. This is the essential problem and why it's not part of the "pay it forward" experiment. The theft was not paying it forward, it was paying it out. It was removing the opportunity for others to be inspired to give.
Take for instance a simple second order impact, someone gets a free coffee of the card and decides that while it was a good coffee, he or she should instead donate 10x that to feed orphans. If more than 10% of people using the card do this, then there's a 1:1 follow on donation effect to the orphans. That is essentially the idea of "pay it forward" as far as I understand it. Sam decided that betting on his fellow humans wasn't a good one (perhaps knowing his own moral failings) and decided to short circuit the process and make decisions for others. This is the essential problem and why it's not part of the "pay it forward" experiment. The theft was not paying it forward, it was paying it out. It was removing the opportunity for others to be inspired to give.