I remember reading the Slate article you're citing when it first came out. It's a fairly amusing one, which can be summarized as: "Wealthy person without a lick of business sense opens the 'café' of his dreams, declines to do any work himself, and goes out of business." It's entirely irrelevant to the economics of the average coffee shop.
First off, it's not clear to me whether the establishment the author opened was a coffee shop or a restaurant. The subtitle says "coffee shop", but the article then talks about finding someone to "whip up pretty crepes". I recognize that the dividing line between restaurant and coffee shop can be a bit fuzzy on occasion, but it's there and relevant. Most restaurants fail.
The really notable part of the article comes at the end, however: "There was, of course, one way to make the cafe viable: It was written into the Golden Rule itself. My wife Lily and I could work there, full-time, save on the payroll, and gerrymander the rest of the budget to allow for lower sales."
Yes, Mr. Michael Idov, you could have done that. You should have done that, from the start. That's what small business owners do: They work. That's what your competitors were doing. The goal may be to eventually make enough money to hire people to do all the work, but that's not where you start. You start by busting your ass.
That's the story of a dilettante who opened a restaurant as a fashion statement and went the way of all such fools.
Anyway, back to coffee shops. I don't know if Mintel's stats include the chains. They probably do. You might note, however, that the article I referenced is entirely dedicated to the proposition that Starbucks's presence in the market has been good for small, independently-owned coffee shops. Furthermore, even if we assume that no chain stores fail at all, a 10% industry failure rate would still have independent coffee shops doing pretty well. (Back of the envelope calculation: 14,000 independent coffee houses in the US in 2005, 11,000 Starbucks in the US in 2010, so let's say indies are only 25% of the total. If 10% of the total fail and all failures are indies, then 40% of indies fail. 60% success rate isn't bad for small businesses.)
And, finally, you're right that the margin on coffee beans doesn't determine the margin of a coffee house. That's what happens when your cost of goods is effectively zero, and I assure you that any restaurant or retail outlet would love for their margin to be solely determined by wages and rent.