With highly molested trouts (my preference), especially on smooth water, my answer is 'all of the above matters'. Here in this part of the state, trouts after the first week are way pickier than those in the Rockies. To concentrate only on the most important factors and overlook the minor ones would be like using a dirty, unwashed glass for wine (even inexpensive ones) - or even water, for that matter. Or listening to a favorite song that's 99.6% fine, but has a scratch in the vinyl (for those oldtimers among you who know what that is).
Regarding the silhouette/profile, I additionally pay attention to the stance of the imitation. The way a CDC/elk (I use deer hair for small ones) sits is much different than the stance of a Trude or Wright Caddis. And an EHC is significantly different than a Colorado King.
Further, many hackled flies dance differently in bumpy water, depending on configuration. Small Dun Variants, with their oversize front hackle, bounce nicely compared to conventional Catskill ties, or for that matter, hackle-free jobbies like Comparaduns. So to stir in a little more complexity to this thick stew of fly tying is the factor of action. And if you are party to a little human-induced motion, rather than trying for the infamous drag-free presentation (even sans microdrag), the design of the fly and its weight and buoyancy, there are plenty of more variations to tie. The Wright 'Sudden Inch' (which actually should be more the Sudden Centimeter, or even Sudden Millimeter) works much better with his eponymous Caddis than with a dubbed flat wing.
Adding shucks and tails or different types changes the surface tension imprint, which has a huge impact at times with glassy water.
Further on the configuration parameter is what portion of the fly is above water vs. submerged. The half-sunken look between the butt-down Klinkhammer is much different than the vertical, Chironomid pupa-like Quigley Cripple.
Related to the color issue is the material response to wetness and floatant. Even at the surface, the hydrophobic nature of fluorocarbons and trilobal enables those tiny microbubbles, which reflect and refract light. Other light effects that are not strictly color but very visual can be obtained from metallics and the plasma treated materials such as certain flavors of Flashabout.
Also the use of monochromatic materials is different than multicolored heathered dubbings. The former is absolutely essential for male Trico spinners, while most other buggies benefit from the latter. The human eye at a couple of feet isn't very discriminating (Seurat's School of Pointillism, anyone?) whereas a trouty eyeball ar an inch or two must discern a bunch of pixels. It gets really interesting when I crossblend different fiber types and gauges.
One common mistake folks make with dry caddis is that the ones flying look a size or several larger in the air compared to in your hand or on the water. And for most of the small ones (ie not the October Caddis(ses) in the PacNW, the body is surprisingly small and often skinny.
A profile/color issue which emerges with the caddis wings is that the trouts are looking up through them. This is significantly different than you looking down at the same wings in your hand. Even the fanatics about color being unimportant should concede that the overall lightness or darkness of the imitation has some bearing on the matter.
For those with neither the time nor inclination to obsess over details like these, stick with wild brookies in small criks or Rocky Mountain trouts (who only have a few months to eat each year and they know it, so they will nip at most anything).
