Adsense, PageRank, and the Way Things Combine

by William Grosso

Related link: http://www.google.com/adsense



I put adsense on my personal website a while back. I did so more out of curiosity than anything else. My site gets very little traffic and it's no big surprise to learn that I can't retire, or even go to the movies, on the advertising revenue from it.


But I've been checking the results anyway. And today, what lept out at me was that Google is tracking the number of visits to various pages with a reasonable degree of accuracy. It's easy for them to do so: adsense is implemented as a piece of javascript that requests an ad from a google server. Every time a page loads from my personal site, Google gets a ping (that's not quite true, but it's close enough). What's more, when search bots and other automated tools hit my site, Google doesn't get a ping (because very few automated tools or spiders execute the javascript on a page). Google is, or could be, getting the equivalent of webserver logs, across a large number of websites.


This is very useful information. Google made their reputation, back in the early days, using PageRank, which they described thusly:



PageRank relies on the uniquely democratic nature of the web by using its vast link structure as an indicator of an individual page's value. In essence, Google interprets a link from page A to page B as a vote, by page A, for page B. But, Google looks at more than the sheer volume of votes, or links a page receives; it also analyzes the page that casts the vote. Votes cast by pages that are themselves "important" weigh more heavily and help to make other pages "important."



The reason this might have worked is simple: the effort involved in creating a link made the vote worth something.
But, for a long time now, the speculation around the web has been that PageRank has been abandoned by google (see also: a google search on 'pagerank is dead'). The most common reason people propose for the alleged death is that the existence of link-farms effectively spammed the results and forced Google to ever more elaborate ways of computing results.


Now suppose that adsense achieves reasonable penetration into the web. What Google will then get is a reasonably good measure of traffic (and something much better than the Alexa Toolbar) for a large percentage of the web. Which will effectively enable it to resuscitate PageRank using people's browsing as a discriminating factor.


What's more, the information that Google gets from adsense is something that competitors won't be able to easily replicate. Adsense not only gives Google revenue, it could give Google an enduring competitive advantage.


Which is a pretty cool side-effect.


Of course, since I don't work at Google and don't really know much about what they do, this might be sheer nonsense. But I think it's an interesting line of thought. What do you think?


2 Comments

gojomo
2004-01-27 22:07:39
Yes, but...
Interesting speculation, especially if...


(1) the ad-insert Javascript can report back the referrer to the AdSense-containing page (I haven't checked if browser JS environments allow this), or
(2) they could somehow catch and report page out-clicks.


Then, anyone in the AdSense program would be donating to Google server-side info almost as rich as what their (and other) browser toolbars collect.


However, I wouldn't agree that his info would be *better* than Alexa/Google/etc. toolbar info. Toolbars can report every click -- by its origin and destination -- and even track page linger times. It seems more reasonable to assume toolbar user provide a representative sample of all web users' activity, than that AdSense sites are representative of all websites' traffic.

wegrosso
2004-01-27 22:22:28
Yes, but...
Well, okay, replace "better" with "different." The toolbars report back on what particular users, typically power-users using Internet Explorer on Windows, see. For example, the browser that comes with Yahoo DSL doesn't let you install the google toolbar. I would be surprised if AOL's environment lets you do so either. It's a very small, and very biased, sample of web traffic. And it's brittle; it requires you to maintain software moving forward. And so on.


Websites using Adsense have a very different, and depending on the total reach of Adsense, more representative, sample bias. They reflect the entire browser population (not just IE on Windows). And they're not nearly as brittle; the burden of maintaining things is lower and shouldered by people who have a cash interest in keeping things going (the webmasters).