In my previous post on testing market demand with facebook ads, I extolled the virtues of the Facebook self-serve ad platform for testing feature/product concepts with live ammunition.

After further experimentation with the platform, I can state confidently that while there are some interesting opportunities for locally oriented products and service owner’s to take advantage of the city-level targeting, there are some fundamental flaws with the way that the Facebook ad auction functions.

The problem lies in the way that the auction for ad impressions/clicks functions. While Advertiser’s bid on a particular keyword in search advertising platforms such as Google’s Adsense, on Facebook, Advertisers bid on a certain demographic segment.

The beauty of Google’s auction is that it creates market clearing bid prices for each and every keyword, pitting advertisers in similar business’s (with similar margins) against one another. On the other side of the coin Facebook’s platform is subject to large distortions in impression/intent fit. This occurs because there is a large discrepancy in the ROI (EPC) that any two advertisers receive from any ONE user demographic segment. Because they are bidding on the same segment, naturally the advertiser with the highest EPC will be able to outbid every other advertiser for that segment for EVERY impression. This eliminates the possibility advertisers with low margin’s to make up on volume (a large percentage of internet advertisers).

Facebook seems to have at least acknowledged this flaw. They try and maximize their own eCPM by lowering/raising bid requirements for each ad dependent on it’s CTR. While this surely optimizes their own revenues, it does nothing to improve the ROI of the advertiser.

Facebook Ad Data: