Before Michael De La Maza published his book "Rapid Chess Improvement", he wrote an article by the same name. Before he published this article, he sent it to me to get my suggestions, because I had a website since around 1996 that advocated the same idea. I believed in doing a large number of tactical problems repeatedly to build pattern recognition. When I did this myself, I saw my rating go from 1800 to 2000.
I can't guarantee that this is the best way to improve one's chess tactics. For myself, my goal was to be able to see the vast majority of 1, 2, and even 3-move combinations instantly. However, doing new puzzles might also be good.
My website hasn't changed much since I first published it. It is dated. At that time, the Internet (World Wide Web) was relatively new, and most people didn't have access to it. Chess Life gave me an "award" for the best new chess website, but there weren't many chess websites at the time.