In other words, a shadow catcher is an object that only "caches" the shadow. In other words let say that you want to render something on top of a photograph, but the photo has an object that the would have to interact with the shadow on something modeled in your seen. You model something to mimic the object on the photo to catch the shadow of your model and make it look more realistic once you do the compositing work on an image editor like Photoshop. It is only meant to distort shadows for compositing work.

The main problem that you might have is that you haven't included a reflection map on your seen. I ran in to this while getting acquainted with Carrara.