The Rank Factor Class taxonomy categorizes snippet optimization hypotheses into alignment, substance, architecture, style, framing, and proof to prevent over-testing single variables and ensure systematic coverage across different prompt engineering strategies.