Efficiency doesn't mean a whole lot for solar. When I say efficiency here I mean "percentage of solar flux converted to electrical power". As solar advocates never tire of pointing out, the total amount of power that falls to earth as sunlight in one day could power today's civilization for a thousand years or whatever.
No, what's really important for solar is cost efficiency, not photo-electric efficiency. The answer: printing solar cells like newspapers. They aren't very efficient, but who cares, they're cheap. Konarka and Nanosolar are two examples. I really don't think their technology will displace coal or natural gas power as the backbone of industrial power. But they'll do a nice job of counteracting my a/c bill in the summer when the sun comes out.
This is the real question, how do we get the cost of panels to come down. 50% of light being converted to electricity is plenty if the cost of the panels is low enough.
With legislation requiring all new buildings to produce a percentage of that building's total usage, say 10 %, we would create a large market for solar panels. Competition would drive the price down to the point that people would begin installing PV panels on their homes because the pay back period would only be a couple years and there after they would actually make money.
There would also have to requirments on electrical providers to install two way metering upon request.
This would be a way to bring PV power on line quickly.