Google Tries To Prove Its Economic Value, but Fails

Last Updated May 26, 2010 6:38 AM EDT

A company like Google (GOOG) that continues to face significant criticism of its activities -- its most recent privacy fiasco is one example -- wants to deflect the ill will. Particularly when many of the harsh words come from politicians and regulators with the power to make the company uncomfortable. That's what Google tried to accomplish in a self-generated report that claimed the company generated $54 billion of economic activity for American businesses, website publishers, and non-profits in 2009. Too bad for Google that a little examination casts a lot of doubt on the conclusions:
  • The company's estimate of the return that advertisers get on their AdWords investment is unimpressive when you look at the details.
  • The methodology for estimating the return is unrealistic in a range of ways.
  • Google avoids saying how much of the claimed $54 billion in economic activity owes to the questionable estimate of benefit to advertisers.
The report rests on two assumptions (I added links to the cited papers):
To estimate the economic impact of Google Search and AdWords, we rely on two conservative assumptions. First, that businesses make an average of $2 in revenue for every $1 they spend on AdWords â€" that's $1 profit. Our Chief Economist, Hal Varian, developed this estimate based on observed costâ€"perâ€"click activity across a large sample of our advertisers; his methodology was published in the American Economic Review in May 2009. Our second assumption is that businesses receive an average of 5 clicks on their search results for every 1 click on their ads. This estimate was developed by academic researchers Bernard Jansen and Amanda Spink based on sample search log data and published in the International Journal of Internet Marketing and Advertising in 2009.
To say that revenue from a marketing campaign less the cost of the ads is profit is incorrect and deceptive. Profit is revenue minus all costs. You can't call the amount margin, because that would mean product revenue less all the cost of goods, and advertising costs wouldn't be part of the equation. In his study, which is an examination of online ad auction pricing, Varian calls the difference "surplus," not profit. He repeatedly cites simplifications of conditions to make the calculations easier. I don't mean that as a criticism, as that's standard when trying to design or use models.

(Varian, oddly, doesn't appear to have played much role in the development of this report, and certainly isn't credited anywhere as an author. When a company like Google goes to the trouble of hiring a "chief economist," though, you can't help but wonder why its big economic-impact report is signed by Claire Hughes Johnson, vice president of global online sales.)

If I'm reading Varian's academic study correctly, a major assumption is that advertisers choose to maximize surplus, which would effectively be the same as maximizing profit. However, is a $2 surplus to ad cost ratio good? Assume that a company sells products at 40 percent margin, which means that the cost of goods is 60 percent of net revenue. Given the surplus figures of $2 for every $1 in ad spending, the total revenue would be $3 and the ad cost would represent an additional 33 percent of the revenue price. Add 60 percent cost-of-goods to 33 percent advertising cost and you're down to 7 percent left. I don't want to call it margin, because, as I mentioned, that shouldn't include marketing cost. It still seems like a poor result.

Varian's calculations also seemed to assume that the advertiser would bid in a rational and optimum manner. Yet, he noted in his paper that advertisers must experiment to determine optimum bids and click-through rates to maximize surplus. Given that the dynamics of consumer response and ad bids are probably fluctuating all the time, many of the advertisers won't be able to work at optimum levels.

For brand-driven advertising, where there isn't necessarily a transaction, value would be even harder to calculate. What of companies testing AdWords? Sophisticated direct marketers would also calculate lifetime value of customers, potentially incurring a net cost to acquire a new customer with the expectation of profit over time -- after the person and company had a relationship outside of AdWords and Google. All this further undercuts Google's estimate.

Then Google says that businesses receive an average of five clicks on search results for companies as ads, "conservatively" estimates that search clicks are 70 percent as commercially valuable as ad clicks, and calculate that advertisers receive a total of eight times in surplus what they spend in AdWords. Ah, what nonsense:

  • Google assumes that people clicking on links are as inclined to purchase -- or whatever brings the "value" -- as people clicking on ads. Presumably, the latter would be more inclined to spend.
  • Any benefit from link clicks has nothing at all to do with having ads. The two are separate events, and a company gets the benefit from search engine optimization and all the work of having a Web site, rapidly increasing the effective cost of using the ads.
  • As the Jansen and Spink study states, "More than 80% of web queries are informational in nature and approximately 10% are transactional, and 10% navigational." (Emphasis theirs.) Therefore, the vast majority of clicks never directly turn into sales, suggesting that the estimate of economic worth of a non-ad click is fallacious.
Now Google discusses the AdSense program, based on the money it receives for ads and the amount it pays out to website publishers, which is a hard number that Google can actually calculate. Finally, it adds grants it gives to non-profits and counts, as the economic value, the amount of the grant. It's an in-kind donation, but counting that as full economic value seems dubious, particularly when it lets Google reduce the amount it would otherwise pay in taxes.

The kicker is that in presenting the information, Google does not state what percentage of the supposed $54 billion is in the actual money paid to web site publishers and what part is the fuzzy "benefit" received by advertisers.

It's not surprising, given that Google sells advertising and those who sell ads are, in general, happy to ... let's call it "stretch the boundaries of perception" on the part of the advertiser. I don't doubt that many companies use AdWords effectively and profitably, but to assume the benefit levels Google claims requires a degree of credulousness that should be difficult for any business manager to summon.

  • Erik Sherman On Twitter» On Facebook»

    Erik Sherman is a widely published writer and editor who also does select ghosting and corporate work. The views expressed in this column belong to Sherman and do not represent the views of CBS Interactive. Follow him on Twitter at @ErikSherman or on Facebook.