Being a graduate student at one of the universities involved, i did some modeling on this event (although we weren't quite up with the game, so our findings weren't used in the report).

The term 'lensing' is a bit of a misnomer, as that implies that you're looking at the source star; which is essentially a giant flashlight that allows us to probe the lens for information about it's planets.

The lens star acts to bend the light from the source, creating multiple and distorted images of it (which are too close together to resolve). Observing the sky from earth, these multiple images have the effect of increasing the net flux measured (in laymans terms, the star gets brighter).
When the lens star has planets (especially, as in this case, one close to what as known as the 'einstein ring') it causes large perturbations to the (otherwise fairly simple) lightcurve. With appropriate mathematical models and massive amounts of computing power, the parameters that give the best fitting theoretical lightcurve can be found.
Combining this with external information and a good dose of physical and statistical insight, it is possible to say to a reasonable degree of confidence (usually never 100%) that you have found such and such a system.

In reality, the astronomers who measure the data are only a very small part of the overall picture, but the media find a much better story in "amateur astronomer finds extrasolar planets" than "scientists use computer grid to minimize 10 dimensional chi^2 hypersurface" so they get all the attention.