More on AdSense 'Smart Pricing'
Salar Kamangar, Director of Product Management at Google, offers some clarification on what factors Google may be using to determine the "expected value of a click" in its new version of "Smart Pricing" for clicks on content-targeted ads.
The general idea, as conveyed in Google's press release, is to differentiate between pages that may attract readers with commercial intent (eg. product comparison) as opposed to pages which are more informational in nature (feature article). Kamangar continues to emphasize the determination of pricing largely "at the page level" but secondarily did mention differential value to advertisers "across sites" as well. Google "analyzes pages based on keywords and concepts to determine which pages are more likely or less likely to convert [to sales]," was another way that Kamangar put it.
I wondered, however, how background data on "expected values" might have been collected in the first place. Was Google running tests with a panel of volunteer advertisers, or was it also analyzing a larger pool of data that Google has access to, the data generated by advertisers using Google's Conversion Tracker tool? "Google looks at all possible pieces of information," responded Kamangar, meaning that he wasn't ruling out use of Conversion Tracker data to assist Google in designing a better pricing model for AdSense.
From an advertiser standpoint, this continues to sound like an improvement that will likely lower prices on some clicks. But advertisers using Conversion Tracker, at least those who have privacy concerns, probably won't find it terribly comforting that their conversion data are fair game for Google to build a pricing scheme that may help competitors (or Google) maximize their profits.
From a publisher standpoint, this seems to take a business relationship that was nebulous in the first place -- Google never published its revenue share percentages, and for all anyone knows, arbitrarily sets different ones for different publishers based on a formula, initial editorial determinations, or something else -- and makes it even more uncertain. How is any given advertiser's inventory being priced? It's anyone's guess. On the upside, the reporting is good, and advertisers can see their average CTR's, CPC's, etc. from day to day.
I asked if data such as PageRank or editorial judgments of quality could also factor into the pricing formula. Again, these weren't being ruled out. "We look at everything we can," maintained Kamangar. "If, for example, we determined that time of day was an important factor in determining the expected value of a click, we might factor that in."
So is Google now using time of day to price clicks? Are advertisers going to find that Google is offering significant discounts for clicks that occur, say, in the middle of the night? "No, not right now," was Kamangar's reply.
Posted by Andrew Goodman
| | Permalink
 |
 |
 |
The Traffick Search Engine Directory :: |
| » Internet Marketing » Internet Tools » Search Engines |
» Web Browsers » Web Portals » Webmaster Tools |
» About the Directory » Add URL » Traffick Report: Flock |
