Battelle links to Greg Linden's latest in a series of thought experiments on Google's vulnerabilities titled, "Kill Google." The thrust of Greg's post is that Microsoft should do whatever it takes, including using Microsoft's "market power," giving away the product for free, and employing other scorched-earth tactics to eliminate Google's AdSense offering. Greg reasons that since AdSense (where Google places ads on third-party websites) is half of Google's revenue, Microsoft should "strangle Google's air supply" rather than trying to beat Google in mundane things like search quality and a better user experience.

But Greg misses a fundamental characteristic of Google's business model -- AdSense isn't as important as it looks. Here's the comment I posted on Greg's blog:

[begin quote]

Even if Microsoft could reduce Google's AdSense revenues to zero, it wouldn't hurt Google as much as you think.

From a financial standpoint, the relevant revenue number is the one *net* of traffic acquisition cost (the amount Google shares with content owners). On this basis, AdSense revenues were only $573 million in 2005, or less than 15% of Google's total net revenues.

That's because Google gives back about eighty cents of every dollar in revenue to the content owner. To count the whole dollar as revenue instead of the 20 cents Google keeps would be the same as eBay counting the total value of merchandise sold on its site instead of just the auction fees eBay charges.

(It's confusing because GAAP accounting rules require Google to count the whole dollar, while they require eBay to count only the auction fee. But economically and from a cash-flow standpoint, the important revenue number is the one ex-TAC.)

For Microsoft to beat Google at the AdSense game would require a lot more than giving the product away - a 100% revenue share wouldn't make publishers as much money as Google's 80% share, since Microsoft doesn't have the inventory of ads to sell or the techology to place them on the right sites. In short, Microsoft doesn't have AdWords, which is where Google gets the inventory of ads to sell.

So the *real* way to kill Google is to kill AdWords, not AdSense. AdWords generates 85% of Google's net revenue, and it provides the inventory for AdSense.

But to kill AdWords, Microsoft has an awful lot of difficult technology to develop, while Google continues to innovate at a higher clock speed than Microsoft. And just as network effects make eBay nearly invincible in the online market for physical goods, network effects make a very high barrier for entry into the market for keyword advertising that Google owns.

Google's certainly not invincible, and I'm looking forward to reading Anil's thoughts on the chinks in Google's armor. But going after AdSense will be a long, hard battle that even if won wouldn't do more than dent Google's cash flow.

[end quote]

I'd posted two years ago about the problem analysts were having with Internet math, and it looks like it continues to confuse people.