• by Mark Simon , January 7, 2008

2007 was a year of extraordinary consolidation in the online ad sector, and one very interesting consequence of this consolidation is the fact that the Big Three search engines (Google, Yahoo, Microsoft) have acquired companies whose suite of tools include Web analytics and online campaign management software. Google (which has maintained its own tool, Google Analytics, for several years) will soon begin to integrate with Double-Click's DART software; Yahoo with Right Media's, and Microsoft with aQuantive's Atlas.

This integration brings with it some very interesting possibilities for the Big Three -- which are each in a struggle to squeeze the most money from their inventories -- and some extraordinary dangers for marketers who use one or more of these tools.

Sharing data through "piggybacked" tracking pixels. Many publishers typically have more than one tracking code on the pages of their Web sites to provide data measuring different aspects of their online marketing efforts. For example, a given publisher might be buying some search media from Google, targeted inventory through Right Media, and perhaps some graphical inventory via DART. To provide a complete picture of the health of these multichannel campaigns, multiple tracking pixels need to be "piggybacked." The obvious problem is that for the first time it now becomes possible for any of these entities to obtain unique insights into how a marketer's given campaigns are doing on engines or properties belonging to their competitors.  

How media prices can be manipulated. Let's imagine a case where Microsoft (using data flowing through Atlas, which has direct access to bid price and conversion data) notices that certain keywords on Google are being bought for $1, but those same terms on MSN are only going for 50 cents. It is very difficult to imagine that Microsoft would not be tempted to raise the rates it charges for these keywords to more closely match Google's. Similarly, Yahoo (using data collected via Right Media's code, including HTTP referrer data) will be able to detect how well certain keywords are performing on the other engines (although it will not have direct access to bid prices). Even without bid price data, it can use this data to perform elasticity tests on high-performing keywords whose purpose is to raise them to what the market will bear through its own network. Google, newly integrated with DART (which has direct access to both conversion and bid price data), will naturally be able to "optimize upward" its own bid prices to provide maximum revenue, and gain important competitive information about campaigns conducted on properties belonging to Microsoft and Yahoo.

But the search engines would never do that! Before you protest "these search engines would never do that!" remember that these are all public companies, and their primary obligation is to their shareholders, not to those who buy media from them. In fact, if the Big Three did NOT use the competitive information they can gain through their tracking pixels, they might well be accused of a fiduciary breach by an irate shareholder. And before you write and tell me "there are laws against collusion," all of these entities will have an excellent defense against a charge of collusive practices: After all, nobody got together in a room and explicitly said "raise the prices so that they match Google's"  -- it was simply the decision of the morally neutral, monetization-maximizing algorithm (which does not have the ability to testify in court) to do this. In fact, because The Big Three are all "black boxes," whose pricing logic is now thoroughly hidden from marketers, it will be practically impossible to prove that the inevitable price rises weren't simply the result of "advances in optimization."

Few parallels in the real world. There are few parallels in the offline world to the kind of situation that is being set up here. The closest analogy that comes to mind is one in which a given business used three different suppliers (say commercial printers) and each time an order was placed with one, the other two were automatically notified as to the price paid. The reason this practice doesn't exist in the real world is that it wouldn't survive for more than five minutes. But for reasons that escape me, nobody in the online world has raised a peep of protest about the kind of world we're heading into.

Who will win and who will lose? The obvious loser in the brave new world of shared media data is the advertiser, who will pay maximum price wherever he seeks to buy media within any of the networks controlled by The Big Three. In effect, in this new world he must now compete not only against competitors within any given search engine or media property, but incredibly, against himself, on another engine or property! So marketers will certainly pay in this new consolidated, data-commingling era. If there's any good news to bear, it's for Yahoo and Microsoft, which will both be able to raise media prices to match Google, hitherto the monetization leader.

What can be done? It's a timeworn principle of business that one should never tell the party one is negotiating with what one is prepared to pay. But there's no easy way around the "insider-data trading" problem. One can do nothing, and simply hope that the engines do not actively engage in monitoring each other's campaigns, but this is a bad bet, given the extreme pressures each is under to better optimize inventory in each network, which means charging more for media. Unfortunately, making a switch from Atlas or DART may be painful for many, especially for agencies that have invested much in training their staffs to use them. In my view, however, the best way avoid insider data trading is to partner with someone "on the outside," which means looking seriously at a truly independent third party who's there to represent your best interests.

advertisement

advertisement