Basing a therapy on joint manipulation is a problem because the manipulation itself creates more joint laxity, and may not actually affect the position of the joint. This has been investigated with MRI and X-ray.
If you've got a postural deficit, which causes a deformation of the vertebral ligaments and altered muscle length-tension relationships (and gradual bone remodeling to boot), which seems more reasonable?
1. Pop them around, which feels good because it stimulates stretch receptors in the soft tissues, but also further deforms them by taking them into a paraphysiologic state.
or
2. Identify the postural problem and attempt to correct, or at least lessen it, so that the individual will be trained out of faulty movement patterns that caused the problem in the first place.
I'll go with the one that tries to put my back in a normal state, without the added flexibility where I don't want it.
Part of the problem is that NO ONE knows why most cases of low-back pain happen, and people want an answer. When alternative therapies promise relief, and seem to offer it in the short-term, it looks good.
Again, MRI studies on the spine and other joints have shown that the degree of joint deformity (from OA or disk herniation or whatever else) doesn't correlate with pain or disability. Look at L5-S1 or the knee with MRI in the whole population, and you'd expect a lot more problems than people actually have.
There are some manipulative therapies that work with certain joint conditions, but they're typically void of intentional popping. These are the type that some PT's employ.