Patrick,
We also forget that on most courses, but not the largely exclusive ones you play, that there needs to be ample tee space.
The industry rule of thumb is 20 feet per thousand rounds, so a 30K course needs 6000 SF of tees, and probably 8000 or more on par 3 holes. That amount of tee space certainly starts to make multiple tees make sense. In woods, we rarely build wide tees to save trees, so the extra space comes length wise, creating different tee lengths.
I agree that on some holes we don't need as much yardage variation as others. Par 3 holes for one, where maintaining 20-25 yard differences between tee markers results in about 55 yards from the front tees. Short par 4's are next.
As to TD's "pandering" remark, I find it hard to think of it as pandering, when you are talking about 60% or so of all golfers, vs the tour pros who will never show up at any given golf course. If we pander to anyone, its that 1% of long hitters who won't show up. As he suggests, the way to do that is to have the guts to avoid the long tee that statistically no one uses anyway.
It still leaves the problem of either targeting your course to a narrow segment of the golfing population, or as you suggest, investing more in multiple LZ hazard creation. An intersting idea, but I think the time tested multiple tee method is probably the cheapest way to cater (not pander!) to each "class" of golfer. It is possible that land, irrigation, etc. costs may change that in the near future, or even now in selected locations. That said, the cost of bunker maintenance is so high, building five of those at various locations down the fw might never be cost effective as five tees and one bunker.