Well, short version: There's not much to explain. Actions only remeber absolute references, meaning absolute layer and file names, absolute tool settings, absolute values. Based on your excample nothing else is shown. The magic of creating the "area" happens all before running the action and from looking at the video it merely seems to use a few things like brush strokes along selections, filters and of course things like expanding the selection and creating more duplicates. It realyl boils down to understanding the tools and settings, not some specific secret that has to do with actually creating the action. So for what it's worth, this is nothing anyone can really tell you. You just have to create your own looks, remeber the steps and then create an action from them - with all the limitations I mentioned. Even loading a different image at some point will require to chnage the action or at least include the pertinent file open dialog using by turning off the suppression of warnings and dialogs, which will make the action slow to run since it runs step by step and of course look much less clever on your part. That said, my impression is that what you want would require soem extra scripting, anyway. So perhaps aim a little less high and split up your actions into sensible chunks? There's nothing wrong with having to select individual layers and then run a simple action on it that represents a sub-set of a larger plan. IMO there's no point in cooking up complex actions that can only create a single look and don't work for all situations. Simple differences in coloration of images for instance will often require additional tweaks, so keeping the coloring part separate from the glittery particles part to me makes a whole lot of sense. You know, what looks great on one image (and a doctored up demo video) can look rubbish on another photo with different lighting conditions.