People will pick this apart and say the study was slanted to provide the best possible results. Here's the thing - there is no single cause for homelessness, and definitely no single solution. This is typically why programs fail - treat the population as a monolith and then cry "failure" when only a small percentage are helped.
They key is proper segmentation and programs/policies that target the groups with solutions designed to benefit them. As a scientific study, this looks extremely successful with this group of people . Now we just have to identify segments within the remainder and test solutions that address their needs.
Obviously this won't get too far with the folks who can't do any critical thinking beyond "Hunter's laptop" or "Lock her up" - but for the serious folks in the room, this study looks like a great start that can benefit an identified slice of the homeless population.