Editor’s note: Last Sunday, we published a guest post by Wharton Professor Eric Clemons on “Why Advertising Is Failing On The Internet.” The post questioned a basic assumption that many of us in the tech industry hold near and dear. It sparked a blogstorm and 600-plus comments, most of them filled with rage. Even Danny Sullivan, the normally unperturbable editor-in-chief of SearchEngineLand, couldn’t believe that Clemons could be serious, and let loose in his own post. He even offered for us to republish it here (and he wasn’t the only one champing at the bit to write a response).
Instead, we invited Sullivan to present a more concise counter-argument, absent some of the raw emotion that fueled his initial response. It is presented below, followed by a rebuttal from Clemons, and then another round. We instructed both to fight clean, but fight hard. In his rebuttal, Clemons offers this startling long bet: “In five years revenues from internet advertising will constitute less than 20% of internet business revenues, excluding revenues from the sale of physical goods. Winner buys the loser lunch and gets to gloat.” Danny counters with his own wager. These two can’t even agree on what to bet on.
Danny Sullivan Believes In The Future Of Online Advertising:
Eric Clemons caused a stir earlier this week with his assertion that advertising will fail on the internet, that “it is going to be smaller, not larger, than it is today.” I disagree. In particular, I disagree with his position that search advertising is “misdirection” where companies like Google are “diverting” customers to places other than where they wish to go unless advertisers pay them. Search advertising, in the way he describes it, sounds like some type of protection racket you might see on The Sopranos.
Clemons provided no proof for the allegations he made against search advertising. He linked to no research nor any studies showing that any of the things he asserted were actually happening. Despite this, he made broad statements such as “misdirection frequently takes the form of charging companies for keywords and threatening to divert their customers to a competitor if they fail to pay adequately for keywords.”
I’ve covered the search space for nearly 15 years now. I’ve covered Google since it first existed. This type of misdirection is what Clemons called “Google’s business model.” But somehow in all my years of covering Google, I’ve never heard of any company being “threatened” in the manner he describes. Not once. If this were commonplace, I think I’d be able to dig up a story or two from over the years that supports the business model he asserts. I cannot.
Certainly there have been lawsuits over the years about Google showing ads that appear in response to terms that are also trademarks. These have produced different rules for different countries. For example, in the United States, you can be prevented from using a trademarked term in the ad copy, but the term can “trigger” an ad to appear. In France, you can be prevented from buying an ad on Google that appears in response to a term that someone has claimed a trademark on.
Somehow, Google has struggled on in France. That it continues to earn money there underscores an important point — not all searches involve brand names. Searches don’t always have the “one answer” that a searcher could potentially be diverted from. Search for “apple,” and what did you want? Apple computers or information about apple the fruit? Search for “fixing a Macbook,” and did you want official advice only from Apple or perhaps third-party resources. If you wanted third-party resources, which among them were the “right” choice that Google might have prevented you from going to by showing an ad? Search for “financial crisis,” and does anyone know which “one” resource was supposedly threatened not to appear unless they bought an ad?
Only a chunk of all searches involve a “one source” answer that would fit into Clemons’s characterization. What percentage that is, I don’t know. I don’t recall stats like this being published, nor did Clemons provide any to base his assumptions on.
Let’s assume it is a large chunk. Let’s further assume that perhaps Google doesn’t need to tell brand owners that they need to buy ads or risk exclusion — they just know that’s how the “game” works. From those assumptions, let’s then look at how it breaks down.
Search for “Yahoo” on Google. No ad, but you get Yahoo at the top of the “free” listings. Search for “American Airlines,” and the same thing happens. So, too, for “Target” or “TechCrunch” or my own site, “Search Engine Land.” Same for “Proctor & Gamble,” “NFL” or “Taco Bell.” Are these companies all ignorant of Google’s shakedown scheme and lucky? Are they carefully crafted pre-positioned exceptions in case this type of debunking was needed?
Search advertising is not misdirection. Search advertisers themselves can attempt to misdirect consumers (and have faced lawsuits and legal actions when they’ve tried). But to say, as Clemons did, “Misdirection, or sending customers to web locations other than the ones for which they are searching. This is Google’s business model,” is a gross mischaracterization that should be obvious for anyone to see.
This also brings things back to his overriding assumption — that internet advertising is failing. Search advertising is the strongest segment of internet advertising, and it has continued to grow. In the midst of the worst economic recession the US has seen in decades, search spending looks to either grow only less slightly than in the past or perhaps have a small single digit contraction. That’s not a failure, not in my book.
Moreover, the downturn in internet advertising seems far more a failure of the economy right now than a failure of existing advertising models. Ironically, while Clemons argues that no one wants to view advertising, there’s plenty of evidence that they will do so in return for things provided for free.
TV viewership is down, yet millions still flocked to watch Battlestar Galactica last weekend because they wanted to see it live — as it happened — and endured the commercials as a result. If they shot over to Hulu to watch it because they lacked a DVR, they again endured commercials in return for the convenience of seeing the recorded show.
Elsewhere on the web, more and more are encountering overlay ads, those ads that appear before you can proceed into a web site. They interrupt the viewing experience. I don’t like seeing them myself. But as a publisher who has used them, I can also tell you they are amazingly effective — nor do they result in mass numbers of people abandoning your web site, where they get good information, free of charge.
Advertising, especially offline, has an issue in that people will avoid interruptions if they can, nor do they particularly trust interruptions. In addition, offline ads are poorly targeted compared to those online and tracking performance is laughable compared to metrics for internet ads. But that advertising still works. On the internet, which continues to grow its audience, advertising is smarter, more targeted, more measurable and ultimately will find a place to be more successful, in my view.
Great response. I disagree with parts of it, but it is well-reasoned and in places quite convincing. Mr. Sullivan presents no data and no studies, but some very persuasive stories. Nicely done.
What is the strongest point of disagreement between us? Mr. Sullivan argues that in all his years thinking through and working through issues in internet advertising he has never heard any company or any individual complain about paid search. In contrast, I have been hearing this complaint from senior vice presidents in travel companies for years, and this year the chorus has been joined by retailers and manufacturers.. I am not suggesting that Mr. Sullivan is deaf, out of touch with reality, or in denial; he just does not work with the same executives I do.
As for others who believe that some types of paid search are a form of misdirection, I suggest that interested readers check out the website for Alliance Against Bait and Click, which is funded by several companies, all of them large household names, all of them major players in their industries, and all of them quite angry. Although it did not pass, Utah House Bill 450 sought to criminalize part of Google’s current business model—specifically the selling of trademarks as keywords to trigger paid search ads. Readers interested in more detail on the subject of abusive search advertising can review the blog of Professor Ben Edelman at Harvard; Professor Edelman has a law degree and takes a more formal legal and regulatory stance in his blog than most readers of TechCrunch might enjoy.
I am convinced that at least some major corporations view sponsored search the way I do, in part because my views were informed by discussions with them. Their principle concern, and mine, is not solely the abuse of corporate trademarks in sponsored search and trigger ads, but the entire nature of sponsored search and Google’s monopoly power in the area. And, of course, Google mostly does not misdirect; it is too clever for that. If most sponsored searches were unresponsive to consumers’ needs, sponsored search would fail.
Google’s business model is to threaten misdirection on sponsored search, but actually to engage in it as infrequently as possible. If the company consumers really want (Marriott, Continental Airlines, 1-800-CONTACTS) pays enough then it is listed in the top lines of sponsored search. If it does not, then Google places a competitor there. Obviously, if the sponsored search links never got consumers what they wanted, sponsored search would not work either for consumers, or, ultimately, for Google. Google does not require that the top companies bid top prices, but it does require that they pay. Google knows who to place on top, but will not do so unless the companies actually bid sufficiently high prices for relevant keywords. This was all treated in considerable detail in my previous post.
All of this really has little to do with the main argument of my original article. Regardless of what Google does or does not do it is not the business model that the rest of the net is going to follow. The New York Times, BusinessWeek,, and LinkedIn are not going to fund their Websites by operating search engines. This should not need to be stated. I am arguing, based on discussions with editors, that they are not going to fund themselves with ads either. This is the essential takeaway of the first half of the post.
So, what is left to discuss? Where else do we disagree? I stand by my earlier points:
The last point, actually, seemed to be the most important. It was really the intent of the article, and the original title was “Business Models for Monetizing the Internet: Surely There Must Be Something Other Than Advertising.” This point got lost in the fury over the title of the article and in rage over the idea that online advertising might lose its importance.
I’d like to offer Danny the following wager: I bet that in five years revenues from internet advertising will constitute less than 20% of internet business revenues, excluding revenues from the sale of physical goods. Winner buys the loser lunch and gets to gloat.
Sullivan’s Rebuttal to the Rebuttal:
Many searches are ambiguous about what should be the “one true answer” that Clemons seems to believe exists. If you search for “is Coke better than Pepsi,” who is wrong for buying an ad that appears in response? Coke, Pepsi or both? Trademark owners have reserved rights to a word, not all uses of it. And in a search for “Are diet colas bad for you,” there’s no trademark usage involved. Who is Google supposed to “threaten” about not getting placement, in that case?
As for the bet, I won’t take that one. That’s because I agree, many sites cannot sustain themselves solely on advertising. Mine certainly doesn’t. Our revenue comes from online ads, paid memberships, lead generation and conference attendance. As a veteran web publisher, I know that in my particular space, online ads alone don’t cover the bills.
There are plenty of other examples. Right now, some newspapers are reconsidering whether they should have “opened” their sites to non-paid subscribers, since ad revenues are plummeting. But even when ad revenues were high, the ads alone weren’t covering all the costs that go into producing the New York Times. Other streams such as print ads and print classifieds were helping to keep the online site going.
But that also reflects the fallacy in Clemons’ argument. In the offline world, do advertising revenues generate more than 20% of total revenues outside the sales of physical goods? I don’t know. I suspect not, and I suspect they never have. Even before the web, the New York Times was not covering its costs 100% through ads. In its pure bricks-and-mortar life, it was using multiple revenue streams.
So yes, I agree that many sites need to diversify their revenue with alternatives to display ads. Some of these methods will be alternative advertising models. Some of them won’t be advertising at all but still generate revenue. Ads certainly haven’t been the sole source of revenues for many web sites in the past, and there’s no reason to think that will change. Similarly, there’s no reason to assume that ads would be the sole source of “real world” revenues for a company.
I also agree that ads cannot be the sole source of funding for the internet. Of course, I never said that this was the case. Nor do I know of anyone who has ever seriously suggested this. Clemons seems to be arguing against an idea few people have had.
Just because many web sites on the web or many companies in the real world don’t earn the majority of their revenue through advertising doesn’t mean advertising is failing. In terms of the internet, I think internet advertising will continue to grow as more people come online and the space continues to mature. And I’m happy to bet this will happen with a far better wager.
Clemons wrote that internet advertising revenues will never be as high again as they are now. I disagree. In five years, let’s look at internet advertising spend reports from some commonly accepted sources, such as the IAB. If advertising is above today’s levels, I win. If not, he wins. The loser pays $1,000 to the charity of the winner’s choice.
Clemons’ Final Reply
I offered Danny what I thought was a pretty clear wager, based on the degree of outrage he expressed in his original post.
It appears that Danny actually agrees. That was the real point of my original article. It looks like we have agreement after all. Internet advertising will be a small percentage of internet revenues. The other business models will become more important. At less than 20% internet advertising must fail as the main support of the internet.
Now we can move on and develop those business models.