Right now I'm sitting in a classroom in Redmond, auditing the UM material so I can be qualified to teach it. I had a thought a little while ago: the $18,500 tuition seems awfully steep, but if you amortize it across the 18 training days, that's a hair more than $1000/day. Divide that by the 11- to 12-hour class days, and you come out with an hourly rate of, say, $90. That's cheaper than hiring many kinds of professionals; there's no way you could hire the caliber of instructors available here (say, Ross Smith IV or Tim McMichael) for that little.
Of course, this estimate ignores the cost of travel, and the time you actually have to invest in the class, but it helps to put the seemingly huge expense in perspective.
There is also lost opportunity cost for the attendee or his employer. Three weeks of billing time can add up as well.
That said, if the cost was $1,000,000 it would still be worth it if the return on that investment justified the cost. Wouldn't it?
As someone going through that training rotation right now, the opportunity cost may seem big in the short term, but long-term, this kind of training -- whether I get the certification now or have to retest later -- is going to make a big difference to my skills. As a result, I'll be able to handle harder projects with better results, which means (in the end) more work and more revenue. Whether or not the consultant raises their billing rate to compensate, there's no doubt that going through this class will increase earning potential.