Members of SWJ often debate the value of metrics and assessments on other threads, but to my knowledge I don't think we have a thread dedicated to the debate/discussion. I'm starting this one based on a discussion/debate that has emerged on SWJ Journal at this link

There are more posts at the site, but I'll start the discussion with a couple of them here:

agree that VSO has been "sickened by the curse of McNamara"- although I'd expand that to the entire DoD has been sickened by it. Rumsfeld's (and Gates' and Hagels') expansion upon McNamara's philosophy: that one can apply reductive analytics to social phenomena (war being a social phenomenon in my mind) has turned us more systematic, not less. The systems demand metrics- and thus every CBA I'm involved in concludes we need more IT systems platforms to better aggregate the data for us- because if only our data systems gave us more simplification of the complexity out there... we'd be golden...
Even if you cannot measure entire complexity of "complex causal relationship", still there are some aspects which are measurable (and are quite useful). Is is possible to measure percentage of villages where I established VSO program (MoP)? It is. Is it possible to determine whether the program is working or not in those villages (MoE)? Sure (Depends if I chose correct criteria). Is it possible to determine whether the insurgents have access to the populace and to what extent? In one village? In some percent of villages in a district? Countrywise? Sure. Is it possible to measure it geographically? Sure. Is it possible to determine the changes in time and space? Yes.

The real problem is not using those measurements per se but using them honestly and correctly.
My response:

You argue you can't plan without metrics which is absolutely ludicrous, unless you're a businessman. Metrics didn't become a major factor in U.S. military campaign planning until McNamara introduced modern industrial practices into DOD. Metrics often have no value based on the fact our leadership will always skew them to fit their preconceived perceptions. Even if we could assess metrics quantitatively or qualitatively without bias (we can't), the results would still mean little. We're not trying to increase our profit margins, the only thing that ultimately matters in a conflict is the adversary's will and means to continue fighting. As long as this persists the conflict will persist regardless of our overly positive or negative metrics. This is what we have seen to forgotten over the past 50 or so years. This doesn't mean we have defeat the adversary via military force, we could reach a political compromise, quit, etc., but regardless of the means our measuring of random and irrelevant metrics means nothing and distracts us from what is important. In fact, I think it can result in creating a distorted reality where we convince ourselves we're winning because we're focused on nonsense that is irrelevant to defeating or acquiescing to the adversary.

If we're fighting the Taliban in Afghanistan, and admittedly that is debatable, then measuring metrics related to economic development, number of villages allegedly pacified by the VSO program, number of troops trained, or even the number of enemy contacts ultimately means very little if the Taliban is still willing to fight. Of course the conflict is more complex than our fight with the Taliban and most recognize that. If we followed Gant's advice I suspect we would have accelerated reinitiating tribal warfare that existed in Afghanistan prior to the arrival of the Taliban by enhancing the means of each tribe to fight each other, and they would play us like a maestro plays a violin.
More rational comments below.