as a rule, selfing a species will produce significantly less variability among the offspring than doing an outcross (i.e. between two unrelated examples of the same species). if you were to pollinate among two divisions of the same orchid, in effect it would be a selfing. species sib crosses between two related but non-identical plants of the same species would tend to fall in the middle from a variability standpoint. when it comes to hybrids, selfing will open up the gates for the full range of genetic variability from its ancestry--as an example, the famous Blc. Goldenzelle was created by crossing the yellow-green Fortune by the classic light lavender Horace. the Goldenzelle progeny share fine form from the Horace, but color ranges from the deepest yellow to an deep rich pink.

species selfing is advantageous if you want to have more plants with a desired trait, i.e. albinism. it's also done when there is what we deem a superior flower, although the vast majority of the offspring will not match the standard set by the parent (which can be considered at one end of the spectrum of variability, the end desired by us humans). the major disadvantage, besides homogeneity, is that vigor is often reduced. outcrosses are much better for increasing vigor, and these days you often see poor-growing albino cultivars crossed to more vigorous but normal-colored plants (with the goal of selfing the F1 and getting more vigorous albinos).