I think that to measure the efficacy of a GM ad based on the same measurable activity that you'd measure an ad for (say) a T shirt company is probably a bit misinformed. No one is checking to see if anyone clicked on GM and subsequently bought a Cadillac.
A lot of the comments on the original article seem similarly misguided. ("I NEVER CLICK ON THE ADS I HARDLY EVER NOTICE THEM.") The McDonalds logo on the back fence of a baseball field isn't to convince you to walk over to it and do something. It's so that later when you drive by McDonalds you "decide" to stop.