Hay Day was, and is, the farming game. Its one perfect interaction is bulk harvesting: swipe and hold across a field and everything gets collected. It feels great. Everyone who touched it wanted it.
I worked on a dragon game that had a little farming off in a corner of the map. The team wanted the Hay Day swipe. The problem is that if you copy a mechanic from the market leader, you've signed up to be as polished as the market leader, and that costs real time and real money for a corner of the game that wasn't the point.
So I asked what the game was actually about. Dragons. Breeding them, feeding them, watching them grow. Bring people closer to the dragon. Toss food and it eats. Level up and it jumps around. Evolve and something genuinely cool plays. Attachment, not throughput.
The game wasn't a big success, so I use it as a case study rather than a victory lap. But the decision was right. The money went into the thing only this game could do instead of into a copy of something another game already did better.
The general version: if a cool interaction works in one game, ask whether it serves your game's job before you port it. Usually the answer is a different interaction you haven't built yet.