One useful data point is the cost of gold from illicit services. The WoW Token has taken a bite from their business, but these people are still out there, botting trade goods and stealing and pillaging accounts despite Blizzard’s continuing attempts to ban cheating accounts and confiscate illegal gold. The going rate for gold from these outfits is about $15 per 100,000 gold.
I've been having a lot of fun throughout World of Warcraft's Legion expansion, but I'm coming from the experience of having played it since the beginning with a heap of high-level characters to romp around with. For many Legion is their first crack at WoW (or their first MMO), and several changes to the leveling system have a lot of people up in arms.
"The way the glyph market works on most realms is that there are a couple of people willing to put in the long hours who will have every available glyph posted profitably, and undercut within minutes of being undercut. They all reduce the price very minimally when undercutting, and the "competition" isn't about price, it's about who can relist more frequently.
Recipe ranks basically make you more efficient at making stuff, either by letting you craft the same number of items for fewer materials or getting a higher yield from the same amount of materials. As an example, having a one-star rank recipe lets you craft items at a relatively inefficient way, but when you acquire two-star or even three-star recipes, it can almost feel like you are crafting items out of thin air because of the low amount of materials required for each craft.
Something to keep in mind as well is that this will make enchanting materials more valuable, since it is likely most players will now simply scrap their boe greens instead of getting them disenchanted. Also redundant dungeon/raid gear can be scrapped rather than traded to the (guild) enchanter in the run. Less stuff gets disenchanted, meaning there will be less ecnhanting materials, meaning their value will rise. All the more reason to DE rather than scrap as an ecnhanter.