There are no easy answers, and certainly implementing any of those is harder still.

One of my beef's with COIN is that it tends to enable the policy types to sit back and wring their hands over how much it costs, how long it will take, how many causualites will be incurred, etc for the military to SOLVE the problem so that they can get back to policy work again. IMHO this is completely, negligently, backwards thinking. All the military can do when it goes out to conduct FID to assist the COIN forces of a foreign ally are to help shape conditions that allow the polciy types to identify, address and repair the failures of local/national governance in the COIN force that created the condtions that are the ripe soil that insurgency grows in (and we all know what fertalizes those seeds, the blood of all swept up in the second order effects of those govenrmental failures), moving the bubble down the curve on my chart from ph 1/2 down toward ph 0, to shape conditions so that the policy guys can actually solve the insurgency.

Right now there should be a giant pair of vice grips on Mr. Karzai's nether regions, with the collective providers of surge forces demanding that he hold a Loya Jirga to address poor governance in Afghanistan. Then, and here is the hard part, those same external policy types need to STFU and allow self-determination to take place, and accept what the people, the leaders, the culture, and the process of Afghanistan comes up with. We must relinquish our efforts to control the outcome if we want to craft an enduring solution for the Afghan populace. Then simply recognize and work with whoever ends up in power. This is populace-centric COIN. Its a strategic concept (One paper here on SWJ, and another in World Politics Review for those interested). Still wrestling with the tactics of 'population-centric COIN". It seems to mean many things to many people.