Authors: Douglas Edwards
"Google is big enough at this point that it's entirely possible the terrorists used it to help plan their attack," he pointed out. "We can try to identify them based on intersecting sets of search queries conducted during the period prior to the hijackings." It made sense. While after 9/11 there were many queries about the World Trade Center and the explosive potential of a fully fueled jetliner, it seemed likely there had not been a great number before that date.
A quick note of explanation about what data was in our logs. We did not have personal information about users (names and addresses), but like most websites, Google placed a unique string of numbers on each user's computer when it connected to our site. This string of numbers is known as a cookie. In our logs, all searches were associated with the cookie and the IP address of the computer conducting the search. So if we saw one cookie connected to several searches relating to the bombing, we might be able to identify the user's Internet address. And by looking at other searches he had conducted, we might be able to determine his real-world identity. For example, if the user had searched on his own name (a relatively common occurrence), that search would be connected by his cookie to the other searches he had done.
Sergey compiled a list of words to look for in the logs, including "Boeing," "aviation school," "Logan airport," and "fuel capacity." In a first run, the logs team found about a hundred thousand queries a day that matched some of his criteria. I added a set of terms I derived from searches across message boards where the names of the hijackers had appeared before 9/11, though I realized it was likely those messages were coincidental posts by unfortunate users sharing the hijackers' names.
As far as I know, no one outside Google had requested that we mine the logs, though news reports indicated that the government had installed Carnivore machines (computers that can monitor Internet communications) at a few ISPs to track conversations across the web in real time. Reports also said the government was searching ISP logs for traffic from a specific email address. The Bush administration's interest in Internet chatter had expanded exponentially overnight.
I had no qualms about helping with Sergey's search effort. No one knew if there were other terrorist cells waiting to attack and if so, where they would strike. If we could provide information that might save lives, we had a moral obligation to do so. The cost in terms of potential loss of privacy seemed negligible, given how constrained our parameters were. We would only try to identify individuals who had displayed, before 9/11, a suspicious interest in topics clearly related to the hijackings.
Still, there was no way to avoid the fact that we were trying to sift out specific users on the basis of their searches. If we found them, we would try to determine their personal information from the data about them in our logs. I think about that when I hear debates over Internet privacy. The "Yada Yada" wording I wrote for the Toolbar had been my first encounter with potential issues of user privacy. This was the second. The debate over user data was not one we had actively engaged internally or with our users, and I worried that it would escalate into a potentially devastating communications issue for us. I resolved to find a way to defuse it—once life returned to some semblance of normality.
The search of our logs for the 9/11 terrorists turned up nothing of interest. The closest we came was a cookie that had searched for both "world trade center" and "Egypt air hijack." If the terrorists
had
used Google to plan their attack, they had done so in a way that we couldn't discover.
I turned my attention to our burgeoning News and Resources page, which had grown kudzu-like into a long list of links to news sources and relief organizations. It had been visited more than four hundred thousand times the day after the attacks. My charge was to keep it current, which quickly became a politically sensitive role. One of our VCs asked us to add a donation site run by a company he backed. One of our salespeople had a client that covered technology. News organizations from around the world beseeched us to be added. I had no set criteria by which to determine who was link-worthy and who was not, so I winged it, checking with Cindy on submissions that I felt could go either way or might have PR ramifications. Users requested news sources in the Middle East and Africa. And Canada. I had neglected our neighbors to the north and they felt under-appreciated. I added the CBC and English-language news services from Arab and African sources. I reminded everyone that the page was a temporary service and "not intended to become a permanent feature of our site." They didn't seem to care.
All week, we walked a fine line through a new set of circumstances, unsure of what our next step should be. It was an instance where we couldn't run user testing or rely on data we had in-house. We had to go with our instincts. Eric forbade unnecessary travel, and we cancelled the launch of a group of international Google sites because Afghanistan was one of the included domains. Eric also cautioned us to be particularly sensitive when interacting with angry users, given the tenor of the times. We did not take his warning lightly. One user, irate about the results returned when he searched for his own name, threatened to show up at our office and "do a
Rainbow 6
on Google's front door."
*
I put our local police in touch with him. Another user, upset about our caching of his copyrighted photos, berated us on the phone with an irrational, profanity-laden tirade. The photos in question were grainy amateur shots of his cat that he had posted online. And so it went. In the days after 9/11, we couldn't write off any threat as coming from just another crackpot.
Meanwhile, I continued to struggle with the tone of our communication to users. On the Monday after 9/11, a Boston University graduate named Alon Cohen emailed us a small image and asked if we would put it on our homepage. It was a red, white, and blue looped ribbon, and it was exactly what I wanted. It wasn't generic, gaudy, or cartoonish, and it didn't shout, "Look at us! We're patriots!" It was simple and tasteful and conveyed respect. I sent it to Karen and asked her opinion. She liked it, so I took it to Larry and Sergey. We posted the ribbon that day—on the homepage—with a link to the condolence message we had put up the week before. Almost immediately an offended user complained because we hadn't used "a real American flag," claiming our ribbon had a "politically correct stench." He was harsher than most, but hardly alone in demanding we display more overt patriotism.
An ad rep in our New York office asked to create Google t-shirts with a stars-and-stripes version of our logo. He wanted to give them to all our clients. His colleagues loved the idea, but it put me once again in the position of saying, "Whoa, Horsey."
"We've been careful with the site itself not to cross the boundary between showing support and calling attention to the fact that we're doing that as a company called Google," I explained in a note to him. "I think this might cross that line by literally wrapping us in the flag." The thought crossed my mind that brand management was all about knowing what you needed to "no." The bigger we grew, the greater the forces buffeting our brand and the more powerful the currents causing us to drift from the anchor of our strategy. Without constant care, the trust we had built with the public would crash against the rocks. I based my decisions on experience, intuition, customer contacts, staff discussions, founder feedback, and my own developing sense of what was "Googley." But I was basically guessing.
On 9/11, the whole world shifted. Old rules no longer applied. Except at Google, where the post-attack anarchy more clearly exposed our normal modus operandi: Larry and Sergey did what they thought was right and the Google brand tagged along for the ride. I ran after our lead-off hitters, always a step slow and a base behind. As soon as I adopted a position I thought they had declared inviolate ("Google only does search"), Larry and Sergey raced on ("Now we do news, too"). I tried pointing us back toward familiar turf by proposing a timeline for phasing out the news page.
"We should return to our normal homepage on Monday, September 24," I recommended. "The longer we wait, the more awkward it will be to remove the ribbon and the link, because when we do, people will say it means we no longer care." The natural flow of news would ebb after the second week, I thought. That was the nature of disasters and the public's attention span. Users would understand our return to business as usual. Besides, we were now crawling news sites on a more frequent basis and including that information in Google searches. We could point people to that service instead.
Larry rejected my plan. Our news search didn't work well enough to use it as a substitute, he said. Sergey believed the United States would attack Afghanistan within a week, and the news page would once again be valuable. Reporters let Cindy know that they loved it, too. And so it stayed, and I went back to tending my little garden of links.
By October 3, it seemed reasonable to start running promotional messages on our homepage again. Promotion lines brought in advertisers and drove use of our services, but we had stopped displaying them on 9/11. If we put the ribbon and the memorial link next to the promotion line, they would look like ads. I recommended that we move them to the bottom of the page, where they would feel slightly asymmetrical and thus temporary, making it easier to remove them in the future. Larry thought it would just make the jobs and corporate info links already at the bottom of the page look 9/11-related too. I was about to respond to his concerns when we started bombing Afghanistan.
Forget the ribbon and the promotion line, Larry commanded. Put up a news headline instead.
"Not so fast," Urs responded. "Are we now a news site? Are we competing with CNN? Why do we have a news headlines on our front page? I can see the point for a one-time event like 9/11, but I don't see the point of doing it now. Can someone explain the rationale?"
The company split on the question. Karen felt it pushed us dangerously close to becoming a portal. Marissa thought we should hold off until we could do news "right" by implementing the automated news service that engineer Krishna Bharat was building. Cindy was still getting smiley-faced comments about the page from her press contacts and wanted to keep it. I argued that portals like Yahoo had dedicated teams that managed news better than Karen and I could do on our own. Our page, to use one of Cindy's favorite phrases that always made me wince, "was Mickey Mouse."
Salar sided with us dissenters. "Do we intend to update the latest news about the war as it goes on?" he asked. "Aren't we setting user expectations and doing a poor job of meeting them?" Salar sounded persuasive, especially when Omid joined our chorus of naysayers. It was a typical Google decision-making episode—input from everywhere thrown into the hopper to be processed by the founders. I tossed in one last point I thought would cinch the deal: only a small percentage of people who saw the news link clicked on it. The use of the valuable homepage space was clearly inefficient.
None of it mattered in the end, because one of the people clicking on the homepage link was Sergey. "As a user, I just want to see what is going on in the world using a few top sites," he told us. "I don't see any rush to get rid of this with U.S. attacks and potential terrorist retaliation."
Sergey found it useful, so it was useful. That was also typical of the way decisions were made. The news link went up and the ribbon came down. For days, then weeks, then months, I cultivated our news page to keep it current—adding links to the Department of Defense, the White House, and breaking news about Pakistan, Afghanistan, anthrax, and the Quran. I checked out African, Asian, and European news reports, the CDC, the FDA, the CIA, and the UN. It made me feel amazingly well informed. I became besotted with my editorial power over a page seen by thousands of people every day. Well, slightly tipsy maybe. It was, after all, just a pile of links, the online equivalent of a mix tape.
One day Cindy asked for the rationale behind which links I accepted and which I rejected. I explained that my decisions were based on the value each site provided to the balance of news already represented. Did it offer a different perspective? Did it reach an audience not already served? How long was its name? Would it fit in the space allocated? Or would it cause my tidy columns to grow raggedy and aesthetically displeasing? In other words, my decisions were completely subjective. Cindy advised me that there were ramifications for the PR team when we left someone out. I needed to be more inclusive and make decisions faster, because reporters who were kept waiting got cranky and their coverage of us might reflect that.
Meanwhile, Krishna had been tinkering with his news-search program, and in November we added a link to a cluster of articles its algorithm selected—the first version of what ultimately became Google News. That day foreshadowed the obsolescence of my hand-picked list of links. Krishna's algorithms could sort much more information, do it much more quickly, and deliver actual stories of relevance instead of pointers to front pages. His breakthrough had the unfortunate side effect of making it harder for newspapers to sell their printed products. My link-list page would be just the first casualty of the automated aggregation of online news. In mid-2003, we took it down for good.
There was one other coda to 9/11. As we approached the first anniversary of the attacks, we addressed again the charged question of a homepage commemoration of the event. Suggestions poured in. "Fly the flag at half-mast off the letter
L,
" a Googler suggested. "Turn the
L
into the Trade Center towers," wrote a user.
Karen, Marissa, Dennis, and I debated a long time before agreeing to keep it simple: we'd put up the same ribbon that we had used before, with the date 9/11/01 beneath it. Nothing more. We'd also only display it to our users in the United States, not those overseas. In preparation for an avalanche of angry email, I drafted responses to the main issues we expected to be raised.