I have some concerns about PPP. At several points in the paper, the author expresses the need for identifying good (nonfraudulent) traffic for the algorithms to work. One could argue that if you could identify nonfraudulent traffic that easily, you'd just use that instead of PPP. Another is that advertisers are primarily interested in the performance of their ads. If advertisers feel they aren't getting the proper amount of exposure, they may seek other alternatives. (Imagine someone paying for an ad to air during the Super Bowl, but there is some probability it will not air during the halftime show.)
As an aside, it's too bad this acronym conflicts with the widely used Point-to-Point Protocol.
BTW, ZDNet is reporting that Google CEO Eric Schmidt believes the "perfect economic solution" is to "let it happen." There have been arguments that as click fraud increases, advertisers correct for it by reducing their ad spend. So he's not wrong for saying this, but the comments may be interpreted (by investors) that Google may not have, or be able to design an effective solution. Investors may "correct" the price that GOOG trades for as a result, the same way that there was a selloff when Google CFO George Reyes implied that Google's business would slow.