To concur:

It's not merely predicated on a utilitarian ethical viewpoint, it assumes that we've accurately accounted for costs of the options (eg, second and third order effects) when deciding rather than merely justifying our preferences with biased models.

Empirically, the second (biased models) happens considerably more often than the first (accurate accounting) -- to the point that even if you're a utilitarian, you have to admit it doesn't work in practice. You simply can't make the required benefit calculations for utilitarianism.

This is something businesses get wrong a lot: their numerical justification is actually a reflection of the biases of their staff, rather than an accurate accounting of the options.