The difference is probably due to how the different business models deal with disengagement. For Zynga drops in profitability are a leading indicator of player disengagement, because players stop being interested enough to fork over money before they totally stop playing. For Blizzard it is the other way around, players often stop playing before they stop paying. I'd also argue that for all it's flaws World of Warcraft is still inherently more fun than any Zynga games. Players burn out a lot faster of Zynga games: Farmville had over 80 million users at it's peak a year and a half ago, and is down to 30 million users today. Zynga keeps being the leading Facebook game producer not by having one game everybody plays (like Blizzard), but by making new games every few months. They are even about to enter the domain of fantasy games with their soon-to-be launched Castleville. And while their games are getting "better", that is from such a low base that players aren't staying beyond the new and shiny, the first dopamine rush. Having to keep players with the company by producing and running so many games is expensive, thus the low profit margin, although each individual game is relatively cheap.
But there is also an alternative explanation for the decline in player numbers for Zynga, Blizzard, and many other players: There are now too damn many games out there! When World of Warcraft came out, it sold more copies on days one than what the analysts had said was the overall cumulative size of the European MMORPG market. Or as Azuriel shows on his blog, WoW basically doubled the overall market size. But very few games manage that feat, and having twice the market size isn't of much help if you have ten times or more the games. Due to their lower barrier to entry, Facebook games suffer much more from this than MMORPGs, there are already over 3,000 games on Facebook. And there are tens of thousands of games on iPhones and Android. Wikipedia write about the North American video game crash of 1983: "There were several reasons for the crash, but the main cause was supersaturation of the market with hundreds of mostly low-quality games which resulted in the loss of consumer confidence." Does this look familiar? History repeats itself.