free html hit counter August 2005 - Page 6 of 8 - John Battelle's Search Blog

First Excerpt

By - August 05, 2005

Thesearch Bookcover-1This post will be the first of a number of excerpts from my book. Over the next month I’ll post as many as I can. This first one is from a chapter I wrote on Bill Gross, who has founded many search companies (his latest is Snap), but Overture (nee GoTo nee Yahoo Search Marketing) was his biggest hit. I think Bill makes for one of the best stories in the book, and I hope I did him justice. This is a small portion of the chapter, titled “A Billion Dollars, One Nickel at a Time.” Each chapter in the book is broken into sections, this excerpt starts with a section, about a third of the way through the chapter, which focuses on Gross’s early insights into market economics. As with all things book related, I look forward to your feedback and clarifications/corrections.

The Sugar Daddy: It’s All About Arbitrage

Gross-TmWhen he was twelve, Gross lived in an apartment building in Encino,

California, outside of Los Angeles. There were hundreds of kids in

that complex, Gross recalls. “We all roller-skated together, played

baseball together, swam together, did everything together,” he tells me.

And when they had saved up enough money, they all made the pilgrimage

to a local pharmacy, where they’d buy their fix of candy.

“We used to hop the cinder-block wall surrounding the complex and

go buy candy for a dime at the West Valley Medical Center,” he recalls.

“We’d go there all the time.”

Now here’s where it gets interesting. In Gross’s words: “One day

I was at Savon [pronounced Save-on] on Ventura Boulevard and saw

they had a special on candy, three for a quarter. So I bought five dol-

lars worth—at eight and a third cents each—and brought them back

to my apartment, where I sold them for nine cents. I saved the kids

a penny, and they didn’t have to hop the wall. Everyone began buy-

ing from me. I would ride my bike up there to get the candy and

bring it back in bulk in a big Styrofoam cooler box I mounted on the


In essence, Gross staked an initial capital investment of five

bucks on an arbitrage opportunity in the local candy market, and it

paid off. He was making two-thirds of a penny on every unit—

roughly an 8 percent margin—but he really started cleaning up as

his volume increased. “After I started buying whole boxes of candy,

Savon sold it to me for seven cents. And then finally, when my vol-

ume got really big, and I was selling at the bus stop and at school in

the mornings, I got it for six-point-four cents, as I recall, from Smart

and Final in Van Nuys.”

Volume had driven Gross’s margin up from 8 percent to more

than 40 percent. With the profits, Gross paid for his next project:

the solar energy kits he sold in the back of Popular Mechanics. “I

made a business in candy that allowed me to buy the math books

and solar energy parts I wanted,” Gross explains. Those kits, in turn,

paid Gross’s way into Caltech.

Gross learned several things from his days as a player in the

candy trading market: first and foremost, it pays to be a supply-side

sugar daddy in the middle of a high-demand transaction with clear

market imbalances. Second, Gross realized that you can make sig-

nificant money on pennies a transaction, if the volume is high

enough. And third, he developed a taste for entrepreneurship, a taste

he has clearly never lost.

What Gross spotted in the frothy search market of 1997–1998

was another arbitrage opportunity. As defined in Webster’s,arbitrage is

“the nearly simultaneous purchase and sale of securities…in differ-

ent markets in order to profit from price discrepancies.” Gross ob-

served that the market for any kind of traffic—be it undifferentiated

or intentional—valued clicks at about five to ten cents each, but it

seemed obvious that the inherent value of intentional traffic should be

far greater. If Gross could harness and sell a search engine’s ability

to turn undifferentiated traffic into intentional traffic, he’d make a

killing on the spread.

But Gross had a conundrum. To launch a search site like, he needed both audience and advertisers—and the more

advertisers the better. (GoTo filled out its search offerings with a

standard organic search feed from Inktomi.) Gross knew he could buy his

audience, and he reasoned he could arbitrage that audience’s inten-

tional traffic—as reflected in the keywords they typed into his en-

gines—against an advertiser’s desire for business. But he needed a crit-

ical mass of keyword-buying advertisers to support his site, and given

the untested and relatively complex nature of what Gross was creat-

ing, it was going to be quite difficult to persuade those advertisers to

come on board and bid for keywords. After all, while Bill Gross un-

derstood the intrinsic value of a keyword, not many others in the In-

ternet world did. Until he could prove otherwise, Gross was selling

theory, and little else.

Gross solved his problem by adopting the time-honored approach of

dumping—or perhaps drug dealing is a better comparison: the first one’s

free (or nearly so). Gross built not one but two entirely audacious

ideas into GoTo’s initial business proposition for advertisers: first

was the concept of a performance-based model—one in which advertisers

paid for a visitor only when a visitor clicked through an ad and onto

the advertisers’ sites. Instead of demanding upfront money from

advertisers, the way AOL or Yahoo did,’s model guaranteed that

advertisers had to pay only when their ads were clicked upon. Of course,

this is now the standard model for the multibillion-dollar paid search


Second, and even more audacious, was how Gross priced his new engine:

one cent per click, an extraordinary discount to the market. He knew his

price was seven to ten times less than what every Internet marketer was

paying at the time, and in an environment where traffic was crack,

advertisers couldn’t help but look to Gross for a fix.

In short, Bill Gross bought traffic from one place for five to ten

cents, and resold it on his site for a penny. Not exactly a great busi-

ness model. But Gross believed that the market would take over,

and that soon advertisers would compete to be listed first for high-

value keywords like “computer,” “camera,” and book titles. On the

come, Gross was betting that market forces and the greater value of

intentional traffic would push per-click prices past his cost of traffic


Gross’s gamble lay in building out GoTo as a habit for both his

advertisers and his audience. Back at the IdeaLab’s headquarters, he

built out elaborate models showing how GoTo would slowly grow

audience and advertiser share, and how his plan of arbitraging traffic

would eventually turn profitable as advertisers began to bid various

keywords up from one cent to as high as two dollars.

“Eventually, with volume, I was able to drive traffic acquisition

costs down to six and sometimes four cents,” Gross recalls. “Then

people would exit paying a penny, or possibly two, as some might

click on more than one link,” he continued, warming to his tale.

“But people were also bookmarking the site, and using it again,

which drove down my average cost to acquire a searcher/search.

With volume and loyalty, my cost to drive a search was declining

each month, and my earnings for each search were increasing.”

In about six months, Gross claims, the two prices met and crossed—the

average price paid by an advertiser rose past the average price GoTo

paid to acquire a searcher. “Our model had them crossing in about two

years,” Gross says, “so we were way ahead of schedule. I was certain we

could get there, because I knew bid prices would increase to their true

value over time, and I knew the true value was somewhere in the [range of]

twenty-five cents per click to two dollars fifty cents per click and even

higher on some terms. I never knew some would go to one hundred dollars [as

they have for terms like “mesothelioma,” a rare cancer that—in a gruesome

twist of capitalist fate—affords a high chance of recovering damages in a

lawsuit], but I was sure they would beat one dollar or two dollars, and

they did.”

Back in 1998, the idea of basing a business on the idea of pay

per click was viewed as a wild and rather dismissable gamble. After

all, if you’re Yahoo or AOL, why would you ever want to be held ac-

countable for the performance of what you sold to your partners? If

marketers couldn’t turn the traffic into profits, that was someone

else’s problem.

“The more I [thought about it], the more I realized that the true

value of the Internet was in its accountability,” Gross tells me.

“Performance guarantees had to be the model for paying for media.”

Gross knew offering virtually risk-free clicks in an overheated

and ravenous market ensured GoTo would takeoff. And while it

would be easy to claim that GoTo worked because of the Internet

bubble’s ouroboros-like hunger for traffic, the company managed to

outlast the bust for one simple reason: it worked.

  • Content Marquee

Google Nixing Cnet?

By -

So says Slashdot and Cnet (scroll to bottom). Say it ain’t so, Google! The story Google is upset about is this one….I remember seeing this piece, which has a bunch of personal info on Eric Schmidt that was found via Google, and thinking “Huh, some of the stuff in that story is old, or out of context,” and thinking that a few more phone calls would have been in order, at the least. But the point of the story, even if it was not necessarily made well, was that so much info was available on Google. And another point might be that a search engine never gives a full or necessarily accurate picture of the person (that point was lost). But to ban Cnet for this? Sounds a bit over the top.

ZDNet: Interview with New Yahoo Search Guru

By -

Dan Farber has a good post about Prabhakar Raghavan, who now runs Yahoo’s research.

From it:

Raghavan, who spent 14 years doing search and data mining-related research at IBM and was lured from his stint as CTO of enterprise search vendor Verity, told me that he intends “to go after the best in world and to get them.”

…Regarding search, Raghavan said, “We have two views of better search. Most people are not interested in search—they want to get things done. The future has to be more friendly to people getting tasks done. You don’t want to spend two weeks of evenings sitting at a keyboard and piecing together a vacation plan. You want a system to go out and find the answers, based on future technology that goes beyond crawling and indexing page.”

Thanks Matt

By - August 04, 2005

Matt McCallister, who recently left IDG for Yahoo, wrote a very thoughtful post on why he couldn’t get The Standard, the company I founded and IDG still owns, relaunched while at IDG. Very true words, especially these:

I wish somebody would launch a media brand that covered the Internet business for people in the Internet business.  I’ve bet my career on this industry, and it would be really nice if there was a brand that stood independently in the middle of it, reported on it with intelligence and depth and integrity, and helped facilitate dialog amongst us all.

Trying Out YPN

By -

Ypn-1Astute readers will have already noticed that I am trying out Yahoo Publisher Network ads on the site. So far, the ads ain’t exactly perfect, but, so far, none of the systems I’ve tested have been. Now, I’m not allowed to direct traffic to them, it breaks Yahoo’s TOS, so I won’t tell you where they are. But if you do notice em, feel free to tell me what you think. (And yeah, I’ve noticed they seem to be cut off. Working on that. I’m not exactly a code jockey….)

Factiva (re)Launches Buzz Service

By -

FactivaBack in 04 Factiva partnered with IBM’s WebFountain to deliver a buzz service of sorts, and I liked the idea. Problem was, the IBM part didn’t quite work right, according to folks I’ve spoken with. Now Factiva is back, this time with a homegrown product (“Factiva Insight”) and a partnership with Intelliseek. The company calls its product an “online reputation management system” and I have to say, without too much cynicism, that I’d wager this one will stick. I sense that many, many companies are interested in where they stand in this sprawling ‘sphere, and they’d trust Factiva to help them sort it out. More soon, I think I am speaking with Clare Hart, Factiva CEO, on Friday.

SEW Sold

By - August 03, 2005

SewAs a pal said to me, Meckler never sells save at the top. Buyer is Incisive Media of the UK (quite possibly the worst use of color on a site in the history of the web). $43 million. My, my. Congrats (I think) to Danny, Gary, Chris, et al.

Yahoo Launches Audio Search

By -

Yahoo keeps rolling with new stuff. Tonight it’s audio search. From the release:

Yahoo! Search is the first major search engine to deliver an audio search product that provides users access to over 50 million audio files from major music services and independent publishers.  Yahoo! Audio Search provides access to a variety of audio files including podcasts, music downloads, albums and spoken word such as newscasts, speeches, and interviews, as well as other audio related information including music videos, album reviews, artist images and artists websites….

… Yahoo! Audio Search’s integration with My Web, Yahoo!’s new social search engine, enables users to save audio searches, creating a personal, searchable Web of favorite audio Web pages. Moreover, My Web users can share their musical tastes with their communities by creating RSS feeds of their saved Web pages.

Gary has a complete review here. From that:

what Yahoo is releasing today is a different. It’s a one-stop service (metasearch, sort of) that allows the user to search, find, and access both “open web” audio files (via a Yahoo crawl) along with audio files from numerous music/audio (fee-based) services including Yahoo’s own Music Unlimited, iTunes, Napster, Rhapsody, Emusic,, and other services. Of course, to download these files you’ll need to pay.

That’s right folks, you’ll need to pay. Get used to this, I’d wager it’s coming from Google as well – search as the interface to finding and then paying for stuff. It’s about time, in my humble opinion. Imagine when Google oneboxes (ie adds results automatically like they do with Local) Google Print, for example, or audio, or video….when nearly every search suggests something that can be purchased. Why, it’s a veritable ecommerce nightmare for Amazon and eBay. No wonder they are so into search…

Dr. Lee Update

By -

Google was clearly thinking long term when it hired Dr. Lee, an AP story shows that they were willing to pay him even if a MSFT non compete made it impossible for him to work for the first year.

The Seattle Post-Intelligencer, joined by The Seattle Times and The Wall Street Journal, filed briefs last week asking the judge not to seal the documents, which included a Google document stating it would continue employing Lee for up to a year and defend him in court if Microsoft wrongfully accused him of breaching the noncompete agreement.

Microsoft spokeswoman Stacy Drake said Google’s employment agreement with Lee “underscores the fact that Google was aware that Dr. Lee had a noncompete agreement with us and had the expectation that by hiring him, he could be breaking that agreement.”

The article also had this great tidbit:

Other documents released Tuesday were largely blacked out, including notes from a May 2004 meeting Lee had with Microsoft Chairman Bill Gates and other executives. The subject: “Kai-Fu Lee’s thoughts on search.” Microsoft redacted most of the words, except for “Google,” “MSN,” and “Longhorn,” the former code-name for the Windows operating system update the company plans to release next year.