A visualizer is worth it when the product changes in ways a buyer needs to see: marble against oak, granite on a floor, a bookmatched wall. It is a poor fit when the catalogue is a handful of stones that one photograph already explains.
The expensive part is rarely the first shaded rectangle. It is the rule set. Which combinations exist, which ones manufacturing will reject, and which ones should change the price. If those rules are still tribal knowledge, the visual comes second.
Performance matters because the tool sits in a sales conversation. A scene that loads late is a brochure, not a configurator. The stone, the finish and the room have to answer while the buyer is still standing there.
Treat the visualizer as part of the product: the same accounts, the same quote, the same admin. A detached microsite tends to drift from the catalogue within a quarter.