Microsoft Research is studying & developing advertising auction systems that are “highly robust to fraud�. One researcher has identified Pay-Per-Percentage as a viable alternative to PPC. This is when the advertiser pays-per-percentage of the impressions instead of a Pay-per-click model.
Pay-Per-Percentage is immune to both click fraud and impression fraud. The challenge is that the ads must be shown in a truly random way, across the percentage of impressions purchased. The paper describes prefix-match: which is a system that is similar to broad-match, “but more compatible with pay-per-percentage”.
No related posts.
{ 4 comments… read them below or add one }
Sure, it would prevent fraud–but it would also prevent performance. This sounds too much like CPM, and there is a reason advertisers (like myself) hate CPM and publishers (like myself) love CPM. There is no incentive to perform on any side, as there is with CPC.
It sounds silly to me. Impressions are easier to fake. The trick to this is that if you’re paying for a % of impressions across many websites, it will be nearly impossibly to know who is inflating pageviews w/ fake impressions. At least w/ PPC you know who sent the click. The other problem w/ paying for impressions: we’ll see a million more “Next” links that were never necessary before.
I recently posted similar views on my blog. I’m in favour of CPA/PPP, however, there are inherent risks associated with distributed ads in this way. Three entries of mine highlight this:
http://www.vinnylingham.com/2006/06/google-launches-cpa-ads-problems-with-the-model.html
http://www.vinnylingham.com/category/internet-strategy/the-search-wars/
http://www.vinnylingham.com/2005/11/future-of-search-engines.html
interesting find. cant say yet whether this concept will be favourable among advertisers. if history repeats itself, any product msft launches via adcenter or in the online media space is likely to bomb.
{ 2 trackbacks }