Testing physical anchors to stop mixing up similar nouns

I've been testing a method where I link words to specific physical locations. I tried associating 現場 (genba) with a construction site I pass daily, which helped me stop confusing it with 演奏 (ensou) during my reviews. Previously, I had a 30% error rate on these abstract pairs, but this 'location-anchor' cut my recall time significantly. Has anyone else tried spatial association for vocabulary?