Never Lost Again: The Google Mapping Revolution That Sparked New Industries and Augmented Our Reality

by Bill Kilday

Keyhole started, barely, in 1999. It ran out of cash, essentially, in 2002. It was resuscitated by CNN and In-Q-Tel (the venture capital arm of the CIA) in 2003. And by 2004, it was bought by another five-year-old company. That company was Google.

I tried to explain the reason for the disappearing logo—the disparate data sets and how sometimes our client software didn’t call back to our servers for a new image. In those rare cases, his logo might not load because EarthViewer did not know that the image being viewed was from Airphoto USA. We only showed his logo when looking at one of his photos. “Desperate databases?” he asked. “No, disparate,” I said, “as in different.” “Why don’t you just fucking say different?”

John asked Phil and the engineering team to re-architect the Keyhole back-end server and front-end client software to allow users to switch between different imagery databases. Soon we had two Earths: one featuring J.R.’s data and one featuring Digital Globe’s. Users could switch back and forth between the two.

He purchased the software, and this kind of subscription sale (at $599 each year) began to add up for Keyhole. Afterward, he said, “Hey, let me ask you something. What if I wanted to take this image out? Could I pay you to do that?” I didn’t know how to respond. I never contemplated someone wanting their property taken out of the database. This was the first time Keyhole was asked this question, but it certainly would not be the last.

This was always good for a quick hit of revenue, but something that could be done only once a month. John asked me to tee up a special price promotion and hit send on that month’s batch of free trials timed to hit people’s inboxes right at the start of the board meeting in our conference room. John’s computer was set up to receive order notifications from our payment processor anytime someone purchased a license. So while John presented the dire cash-flow situation to the board, his presentation was frequently interrupted by notifications of orders for a Keyhole subscription. Now, I won’t say that this little stunt saved the company, but John did smile wryly after the meeting and said with a wink, “It certainly didn’t hurt the mood of the room.”

A sliding scale was created that would award equity to people who agreed to a temporary reduction in pay. It ranged from 10 percent to 100 percent. Most of the company agreed to take a significant cut. Some employees, like Chikai, were willing to forgo their salaries entirely in exchange for more equity. Expense reports remained unpaid. Countless calls from J.R. and other various vendors were avoided. We had a plan for one more quarter and we hoped that one of those large deals with CNN or In-Q-Tel would come through. We had three more months before John would have to shut down the whole idea.

“I got them to agree to put our URL up on the screen anytime they use our software on air,” he added. I acted excited for John’s benefit. He had been persuaded to accept the lower price in exchange for this on-screen attribution by a former MBA classmate, Omar Téllez. Both Omar and John had young families and socialized together often. After a dinner at John’s house, when he was describing Keyhole’s struggle and the challenge of closing CNN, Omar suggested that they simply get on-air attribution. Later that evening, CNN’s twenty-four-hour coverage of the U.S. invasion of Iraq opened a new segment at eight o’clock, with Wolf Blitzer bursting onto millions of television screens.

“I’ll try rebooting the auth server again,” Ed said to Chikai. Ed had kicked this can down the road before. While he had raised the point that the authentication server (auth server) code base was in need of a complete rewrite, it had never risen to the level of importance to get done, due to limited resources at Keyhole.

“Guys, we’re down,” said Norm McClendon, the burly yet soft-spoken head of technical support for Keyhole and also Brian’s brother. He had joined Ed and Chikai in the cramped, overheated server room, which was more closet than room. Norm’s support lines were ringing endlessly as Ed tried for the third time to reboot the auth server. Each time it restarted, however, the crush of traffic overwhelmed the cobbled-together machine.

That afternoon John told me that Rob Painter, a partner with In-Q-Tel, had managed to get through to him on his cell phone. After months of bureaucratic back-and-forth, the In-Q-Tel deal was finally signed. At $1.5 million, it was by far our largest contract to date and enough to fund Keyhole’s operations for six more months. Those who had dropped their salaries received makeup payments and everyone returned to their original salary levels. With a maturing product, growing revenues, and working capital, Keyhole was finally on solid ground. An aggressive timetable of project deliverables was a part of the contract.

ABC News was the next to go online, and Peter Jennings began using Keyhole EarthViewer. Al Jazeera signed on. The BBC signed on. CBS expanded its usage, including 60 Minutes. The broadcast business opened several other avenues—from an ESPN bass fishing show to the Tour de France to local news affiliates all over the world. The broadcast business ended up being an unplanned boon to the company.

Later that afternoon, John left the conference to take a work call, and I walked around, carrying my box of dwindling T-shirts and CDs, and encountered another demo with equally dubious commercial prospects. The company’s founder, a friendly Chinese national, manned a small table at the rear of the exhibition hall. Unlike most participants at the conference, he was excited to meet someone from Keyhole and to show me what he was working on. A Keyhole subscriber, he opened up EarthViewer and flew down to San Jose, California. He had a folder of KML placemarks represented as dots along a boulevard in downtown San Jose. Then he showed me that if you clicked on a placemark dot on the map, an info bubble window popped open, and inside of that window was a picture. He told me that it was a photo of that exact location on the map taken from street level. I was amazed: I had never seen such a view. Also inside the pop-up window, next to the photo, arrow icons pointed in different directions. He explained: “You can click on the arrows and it shows you a picture of the street in four different directions.” The experience allowed you to fly down to street level in Keyhole EarthViewer, click on an icon on the map, and then turn in any direction to see the view of that location. “How did you get those pictures?” I asked. “I mounted four cameras with GPS on top of my car.” He pulled up a photo on his computer of his Chevy SUV with what looked like a disassembled mutant cyborg lawn mower ...

Right away, it became apparent that this was one of the most economically ridiculous business ideas I had ever heard.

I hired an intern named Jason Cain to import the digital versions of transit maps throughout the United States. By this time, these maps were being shared as KML files, which Jason could import into Keyhole. The city agencies were contacting us and saying, “Hey, we have this data.” It was the first time that I fully understood what John Rohlf was talking about during our brief exchange in the bathroom.

But Monday came. And Monday went. With no announcement from John. In fact, John was holed up in a conference room with Brian, Michael, and Noah. As I walked to the kitchen or bathroom, I fleetingly glanced through the porthole window of the conference room door and saw them still gathered around the table. When I left work at 7:30 p.m., there were no signs that they were going to finish up any time soon. I, like many others, began to worry. Had something gone wrong with the deal? When I arrived the next day, it looked like the four of them had never left, as they were already back in the conference room. Finally John emerged late that Tuesday afternoon. He looked relieved. I got the sense that something was up and commented on the marathon-like nature of the meeting. In response, John asked me if I wanted to grab a beer after work. We both drove to the Sports Page, parking next to each other in the lot, which was unusually crowded for a Tuesday. I was surprised that John was waiting for me in the parking lot because we often met inside the bar. As I got out of my car, John stood at the rear bumper of my car, blocking my path to the bar’s front door. His expression was stern. “Hey, listen, I need to tell you something.” Looking around, John continued: “You’ve got to promise not to tell anyone. Not even Shelley.” “Okay.” John stood there in the middle of the black asphalt parking lot. The sun was setting against the pristine blue sky. He paused as another group of strangers ...

As I sat there, still stunned, I formulated a list of questions, but one was paramount in my mind: Why the hell would Google want to buy Keyhole? Google didn’t do maps. It had no mapping product. Zero. Zilch. Nada.

Two weeks earlier, a dozen executives met at Google for a product review of the company’s photo-editing software, Picasa. Midway through the meeting, Sergey Brin, the brilliant Russian engineer, flip-flopped into the meeting fresh off the sand volleyball court at the center of the Googleplex. Opening up his laptop, he soon became distracted by something Google engineering head Jeff Huber, a good friend of Brian McClendon’s, had sent him. The Picasa product manager was a few slides into his presentation, but began to notice that others were distracted by what was on Sergey’s screen, too. Eric Schmidt stopped the meeting and asked Sergey if he cared to share what was so important. With that, Sergey jumped up, swiped the projector from the jilted product manager, and started projecting what was on his laptop: Keyhole EarthViewer.

At all of these introductory meetings, I was amazed at how little anyone talked about pricing and sales and revenue forecasts. It was a foreign concept to many at Google that we charged for our product. In fact, generating revenue seemed to be of little importance among its various teams. Instead, the Googlers were focused on ambitions like user delight and changing the world with technology. Making money was rarely, if ever, brought up.

“Just ship it” was a common mantra at Google in those days. There was a strong desire to be nimble and fast, and to not let perfection stand in the way of progress.

“You have to understand,” he said, “a lot of what we went through was not all that pleasant for me. The delayed payrolls, the lawsuit, the late vendor payments at Keyhole. And then the turf wars with Bret and Marissa, and product battles internally at Google. The long hours. It was hard on me. It was hard on Holly. It’s hard for me to put myself back there. Plus, I’d rather think about the future, and what’s next. What’s coming. Not the past.”