William,
I think that's a great question. Simply put, they currently provide different pieces of information, and which one you need more is probably individual. The classic metaphor used is that 2D video is like an x-ray and 3D is like an MRI. If you want to see if there's a broken bone, you want an X-Ray. If you want to see if there is a ligament tear, you'll want an MRI. In both cases, you have a problem you want to solve and a person who will gain from the extra information.
3D by itself doesn't give the answers, just like 2D by itself doesn't currently give the answers, but it can help provide better information for someone trying to solve the problem. Both could be a problem. Some coaches chase a perfect graph just like how coaches used to chase a look like being on plane all the time.
3D
I think 3D is helpful in identifying a global pattern tendency. It's great for seeing the timeline of which movements happened first. It's great for seeing small patterns of acceleration and lack of stability. It's great for quantifying power sources. It's great for showing sequencing and consistency markers. On the downside, it doesn't show the quality of movement (only quantity). It doesn't show some of the most important control centers (no one currently has a great spine model or shows enough detail of how the shoulders/shoulder blades are moving).
2D
I think 2D is better for seeing the quality of movement. It's great for identifying the big contributing factors to a specific problem (low point, direction issues, etc). Currently, it's better for looking at the shoulders/shoulder blades, the spine, and the feet (which can be really useful from a coaching standpoint).
Ultimately, it comes down to the coach and how they think the information will help them.