The article attacks measurement based approaches, in general, which I don't think is supported by the anecdote of McNamara's failings in Vietnam.
Kill count was simply a terrible metric of victory.
Vietnam was a civil war, which requires a political victory. This means allegiance of local leaders and populist support – both of which are measurable values. A kill count is often inversely correlated with political support.
McNamara's failing here is more like the XY problem [1]. He incorrectly assumed that a brutal display of military power was the way to political victory and so he measured the wrong value.
Kill count was simply a terrible metric of victory.
Vietnam was a civil war, which requires a political victory. This means allegiance of local leaders and populist support – both of which are measurable values. A kill count is often inversely correlated with political support.
McNamara's failing here is more like the XY problem [1]. He incorrectly assumed that a brutal display of military power was the way to political victory and so he measured the wrong value.
[1] - https://en.wikipedia.org/wiki/XY_problem