1) The procedure puts an absolute stop to willingly busting minima/stabilization criteria by PIC under older-generation CRM environments (with steep cockpit gradients and more). Hopefully, superfluous now.
2) On paper it looks much worse than it actually is.
3) It requires a positive decision to proceed towards the landing. Very different from the required positive decision to G/A, which is not easy to build under all circumstances.
Pages 19-20 here [year 2000] and
Page 48 here [year 2018]. You don't need to be go-around minded, the procedure is.