Tuesday, April 26, 2011

"The TGIF Show" from 1999

I finally got around to putting up the final bit of video I shot shortly after I started at Google in 1999. Watching it, you'll get a feel for the unstructured nature of our weekly all-staff TGIF meetings back when the company had fewer than 60 employees. Ad libbed remarks. New hire introductions. The board of directors presentation. Lame jokes. Off-key singing. Birthday cake. Silly string.

Many of those elements were still in place when my boss gifted me with responsibility for running TGIF shortly after Valentine's Day in 2003. Larry and Sergey wanted some changes made, and my boss thought it would be good project for me to take on. In my next post, I'll talk about how TGIF evolved into a massive multi-media production as it came to consume ever increasing chunks of my life.

In the meantime, please, watch your head as you duck into the TGIF time machine.

Wednesday, April 20, 2011

So different, yet so alike

I ran into an administrator for the Knight Fellows program tonight at a book signing party for Steven Levy. We shared notes about the parallels between journalists and engineers and it reminded me that I had once put together this handy comparison chart describing Google and the San Jose Mercury News when I made the leap from the latter to the former in 1999. While the differences were great, there were some surprising similarities.

The Mercury News (1999) Google (1999)
>150 years as a profitable operation <1 year as a non-profitable operation
Thousands of empoyees, many with more than 50 years of service Tens of employees, most with fewer than 3 months of service
Well-defined roles, with 7 unions to enforce them No defined roles and strange looks if you ask about them
Key decisions made around an imposing boardroom table by a committee with the publisher presiding Key decisions made in the cafeteria line while a founder is loading his plate with baked organic tofu
All new products based on P&L projections for five years out Most new products based on an engineer developing something Larry or Sergey thinks is cool
Products not released until perfect - this is the first draft of history and the newspaper cannot appear fallible Products released as soon as they're checked for security and stability. We'll let users tell us what needs to be fixed
Smart, articulate journalists, who know what people really need, even if they don't Smart, articulate engineers, who know what people really need, even if they don't
No tolerance for marketing, which is an unfortunate necessity, but taints the journalistic mission No tolerance for marketing, which is an unfortunate necessity, but taints the engineering mission

Tuesday, April 12, 2011

What's the opposite of porn?

I got together with a group of Xooglers recently and heard a story that was new to me. I found it pretty amusing.

At one point Google was refining its porn filter and and came up with a list of terms that gave a strong indication that a web page contained sexually explicit material. If these words appeared in any combination, the page was likely full of adult content and should be screened out.

There was one word on the list however, that had a very high negative correlation. That is, if this word appeared on a page in combination with a known porn flag, the page should still be considered safe to view for everyone.

That word was "county."

I didn't understand why this would be true until one of the Xooglers explained. Without making "county" a negative indicator, he told me, Google users would have a very hard time finding information about Beaver County, Pennsylvania.

Google has likely developed far more sophisticated filtering techniques since then, but I won't be surprised to hear that adult sites are adding the word "county" to their pages as a good luck talisman for the next few months. If it actually works, I owe spam-meister Matt Cutts an apology and a beer -- perhaps a nice I.Porter.A.

Sunday, April 10, 2011

A public disagreement

I'm about halfway through Steven Levy's excellent book on Google, "In the Plex," and enjoying it very much. He obviously did his homework and he drills down into topics I skipped in IFL, like the sunsetting of our premium ads system and the development of PHIL, the artificial intelligence system behind AdSense. I'm learning things I didn't know, even though I worked at Google at the time. There are, however, a few areas Steven and I both glossed over that I can expand on a bit. For example, on page 76, Steven mentions Larry and Sergey talked about underwriting programs on NPR in 1999, and concludes, "...thus began a long history of public radio sponsorship." Well, yes, but there's more to the story.

Public broadcasting was an obvious promotional choice for Google. Larry and Sergey liked supporting worthwhile endeavors related to science and information. On the other hand they believed paid advertising was (a) ineffective, (b) unmeasurable, and (c) obnoxious. The felt the same about sponsorship of sporting events, which didn't stop the calls from event producers, TV salespeople, and pro and semi-pro athletes in every area from surfing to soccer. The most persistent calls came from one particular group.

"Hiya Doug," the drawling calls would begin. "I hear you folks have a real fast search engine over there. Ya'll know, we have some pretty fast engines over here too . I think Google would be a perfect match for BillyBobJimDale's NASCAR team. Think how you all's logo would look moving around the track on the hood of his car at 200 M.P.H."

I didn't bother asking Larry if he wanted to consider it. His idea of an automotive sponsorship was to give $10,000 to the Stanford Solar Car development project. They pasted a barely visible Google logo decal on the side of their rolling science project. Later, we broadly expanded our automotive sponsorship program to include Carnegie Mellon's robot car entry in the 2004 DARPA Grand challenge. That got Larry thinking about new product possibilities that might someday make NASCAR obsolete.

Just as we didn't do sports, we didn't buy TV ads. No Superbowl spot for us. But tasteful announcements on Morning Edition or a nice logo on NOVA would not only let people know we were hiring engineers, but would reinforce the idea that Google -- and its founders -- possessed an erudite intellectual attitude toward marketing. I liked that idea from a branding standpoint and included it in my first marketing plan, because it let us distinguish ourselves from our competition by NOT spending money on the traditional venues companies used to garner attention. And it made it easy for me to just say "no" instead of enduring longwinded pitches and powerpoint presentations about hospitality tents on the 6th tee at Pebble Beach.

I was the point person for all our contacts with public broadcasting, because having spent more than five years at San Francisco public radio/TV station KQED, I had some experience in how member-supported media worked. It didn't work the way Larry and Sergey expected, which led to frustration for them, for me, and for the underwriting executives with whom I was negotiating.

I began talking with NPR about sponsorship in early 2000. They proposed a package around their highest rated shows: Morning Edition, All Things Considered, and Car Talk. The problems began with finding wording that would pass their review process. They wouldn't let us mention we crawled a billion web pages because according to their research, that was half of the entire Internet and thus too bold a claim -- it would make us stand out from other search engines by comparison and comparative language in underwriting credits was verboten. Okay. Well, at least we could have some fun with our Car Talk language right? Something like, "Google. A fast, clean engine for online searching. Never needs a tune up."

No. No comparative language also meant we couldn't claim to be fast, accurate, or relevant. And since we had asked, no we couldn't include glowing testimonial quotes from newspapers or magazines.

We could say something dry and lifeless like, "Google, a search engine for finding information on the Internet from computers and wireless devices. G-o-o-g-l-e.com." We ended up with a variation of that, but no one at Google was thrilled with it. Larry and Sergey thought the rules regarding acceptable language were unduly restrictive and Larry mentioned that to NPR president Kevin Klose when he came by for a visit in 2003. Klose was interested in learning how NPR might take better advantage of new technology and wanted to encourage Google to dedicate resources, if not cash, to help them embrace the future. I was in the room as Larry gave him an earful.

What NPR should do, Larry told Klose, was make it possible to start listening to a program in the house, pause it as you walked out the door, and then start listening again from the same point when you got in your car. And they should make it easy to accelerate playback of broadcasts. It was painful to listen to all the long, thoughtful pauses during their programming. Larry recorded the programs he liked and played them back at double speed and thus improved the efficiency of his information intake by a factor of two. I don't know what Klose had expected to hear from us, but I don't think it was that.

By that point in Google's history, our hunger for engineering talent had become all-consuming. To figure out where we might find talent like that already inhabiting our cube farms, I surveyed our engineering employees, asking what media outlets they tuned to. Unsurprisingly public radio news programing was high on the list. More than half of the engineers listened to Morning Edition or All Things Considered. Third on the list with 36% was the NPR quiz show, "Wait, Wait, Don't Tell Me." (WWDTM ). So it seemed like a match made in heaven when NPR offered us a sweet deal to become exclusive sponsor of that show. I loved the idea. WWDTM was quick, funny, and news-based and NPR told me the show's producers used Google to research their quiz questions.

Unfortunately, Larry was not a fan. "I just think the information content is low," he told me. Well, yeah. It was a quiz show, not the BBC World Service. More surprising to me was that Sergey agreed. He didn't listen to NPR for entertainment, just news. I argued that the empirical data from the survey made it clear we should move ahead anyway, but even in a data-driven company, it was the founders' prerogative to ignore data they didn't happen to like. It was a great opportunity that I had to turn down.

Soon after, Sergey had me go after another great public broadcasting opportunity, but that one turned us down. Our founders were big fans of NOVA, the PBS science show. It was content rich and very much aligned with our brand as a company that relied on science to improve its products. I asked a PBS rep how much it would cost to be one of the three national sponsors of NOVA. The cost was $2.3 million and they happened to have one slot left. I told Sergey.

"So NOVA gets $2.3 million in sponsorship, divided among the three sponsors?" he asked me.

"No. It's $2.3 million each." That might have been a deal breaker, but it turned out the deal was already broken. The other two sponsors already in place were Sprint and Microsoft. When Microsoft heard we were asking about sponsorship, they exercised their right to block us because we were a competitor. They had been blocked previously by another company and now they were doing it to us. That didn't sit well with Sergey.

"There seems to be something rotten in the notion that Microsoft could block our sponsorship," he complained. It was anti-democratic and didn't fit his notion of the public part of public broadcasting. Besides, in 2003 we weren't directly competing with Microsoft. Okay, maybe a little with MSN search. Sergey wondered if maybe we should push back. The PBS bureaucrats might think it was OK, but what about the viewers of NOVA, or the show's producers? Or, for that matter, the decision makers on the congressional committee allocating funds to PBS? Sergey didn't like to be thwarted by policies and decisions made by people. If something was physically possible, we should be able to bring it about with adequate application of intelligence and innovation. We never made a federal case of it, but Sergey remained steamed at Big Bird's nest featherers for a long time.

Fortunately, KQED came through. We might not be able to secure a spot on the national NOVA broadcast, but KQED was happy to put our spots on in San Francisco during their allotment of promos around the program. They ignored the threatening noises emanating from PBS, which was itself feeling heat from Microsoft. That endeared our local public station to us and secured a spot for them in our limited ad budget for as long as I was at the company. Given the machinations in D.C. around funding public media, I hope Google ups its financial commitment to our local PBS/NPR affiliate. The station braved the wrath of Redmond and its own network overseers. It would be a nice to repay them with a very public gesture of support.

Saturday, April 09, 2011

A photo of the pre-Plex

I've previously posted some video of the Plex as it appeared shortly after I joined Google in 1999. Below is a photo taken by our first operations manager Jim Reese from an even earlier period in 1999. It shows Google immediately after the move from the University Avenue office in Palo Alto to an office park in Mountain View.

In the beginning we all had wood doors and sawhorses for desks. Why? According to George Salah, our facilities manager at the time, it was because "They were cheap and easy to get. Larry and Sergey were proud of the fact that we went to Home Depot and spent $19 on these sawhorse legs and $125 on the wood doors. They didn’t even put any varnish on them. It was these unfinished doors and they just threw them on the sawhorses and barely even put any screws in some of them. They were falling off. And that worked for a while and it was great when you had these big massive CRT monitors that were heavy and you needed all this space on your desk for multiple monitors."

In the photo the office looks relatively uncluttered. That didn't last long. When George joined those welcoming Eric Schmidt to Google as the new CEO in 2001, Eric pulled him aside. “Do me a favor,” he said to George. “Clean this place up a bit. There’s trash everywhere, broken bikes and toys and all kinds of stuff.” So George did. He brought his team in one night and swept up all the crap. The next morning Larry sent him an email.

“Where did all my junk go? I want you to bring it back NOW.”  Fortunately, nothing had made it as far as the dump and soon the halls were back to their customary state of disarray.

Eric was a quick learner. Six months later he pulled George aside again, closed the door, looked him in the eye, shook his finger and said,  “I don’t want you screwing this up.” “He wasn’t scolding me for a specific action,” George explained to me, “he was just warning me not to change things, to keep things the way they are.” Clutter might be a sign of a disorganized mind, but at Google it was often just an indication that people were too busy to clean up after themselves. That was not something Eric had any interest in changing.

Monday, April 04, 2011

Whose shot? J.R.'s.

Jonathan "J.R." Rosenberg, Google's vice-president of product management announced today that he's leaving the company after nine years. I remember the first time I met him very clearly. Jonathan makes a strong impression. I wrote about it for IFL, but had to cut the section for length. In honor of Jonathan's departure from the Plex, I offer it here.

It was January, 2000. Larry and Sergey had reluctantly agreed to the board of directors' demand that they bring in a high-powered leader to establish a product marketing group. No traditional consumer marketing person could possibly impress them, though. Jonathan, an executive at Excite@Home, fit the bill. He could talk tech with the best of them. As his potential subordinates, my marketing colleagues and I were asked to give our feedback. Jonathan had already gone through two interview rounds, so we knew which way the wind was blowing.

At the appointed time, I awaited Jonathan’s arrival in one of our conference rooms. “Once more unto the breach, dear friends, once more!” Jonathan’s voice preceded him as he sailed into the room on a gust of enthusiasm. “Or close the wall up with our English dead! Henry the Fifth, act III, scene one. Would you like to hear the rest?”

For the next hour, Jonathan put on a dazzling performance that included a torrent of words about product development, marketing, and the complete works of Shakespeare. Jonathan educated me on brand-building and statistics. He sprayed the whiteboard with pictures and graphs and circled the room as if tacking against some unseen breeze. His voice rose and fell as he rode over questions about working with creative teams, building collaboration and recognizing contributors. He left me awash in information, but skeptical about whether he saw subordinates as a crew of contributors or as galley slaves pulling to the cadence of his commands.

My coworkers and I agreed that Jonathan unquestionably had the intellectual chops for the job, but that our nascent corporate marketing group would be swamped in his wake. The company was small enough that Larry and Sergey apparently felt an obligation to acknowledge our concerns, without bowing to them. They offered me a chance to chat with Jonathan again in a more casual setting.

The Mountain View In N’ Out Burger is an unpretentious paradise for fans of grilled cow flesh. I’d never been there before, but once we arrived, I understood why Jonathan had chosen it. He spoke the local lingo. “A 3 by 3 is three slabs of meat and three slices of cheese. You can get as many layers as you want. Protein style means a burger with no bun and a veggie burger is a bun with no burger…”. There was no avoiding edification in Jonathan’s company.

Once seated on the ketchup-covered benches, we talked about family and work and balancing both in the Valley and by the time I’d wiped away the residue of my Animal Burger, I’d decided he was not the Great Satan I had feared. I sighed. I could work with him, which was lucky for me, because one way or another, Google would have a product management organization. Our marketing group would undoubtedly be subsumed. I would learn from Jonathan, whether I wanted to or not, as we adopted a top-down structure and clearly defined roles. I braced for the changes, telling myself, “That’s how companies grow.”

A couple of days later, Jonathan did the one thing I did not expect. He told us he would stay at Excite. It would be two years before he once more showed up at Google to claim the role of Vice President of Product Management. J.R. had passed when offered his first chance at Google, but Google's execs kept him in their sights and eventually nailed him. The question was, once they had him, what would they do with him?


Friday, April 01, 2011

The Once and Future King

On Monday, Larry Page takes over as Google's CEO. Again. Larry was CEO for the first couple of years that I worked at Google, before the VCs won their battle to have a more experienced manager front the company and Eric Schmidt was brought on board. There are some interesting insights floating around the ether about Larry and the impact of his leadership on life in the Plex. I'll be sharing a few of my own thoughts on this Monday on Bloomberg TV, but here's a preview.

One of Larry's last acts as CEO before Eric took the reins in 2001 was a reorganization of Google's engineering group. It was painful for all involved. I give the details in IFL, but essentially,  most of the project managers who were overseeing the engineering group - giving performance reviews, maintaining timelines, protecting the engineers from random requests (including Larry's) - were not themselves technical specialists. Larry couldn't abide that, especially because they kept interfering with his ability to push through his own large-scale projects, including scanning every book in the world. So he held a full-engineering meeting and told the project managers in front of their colleagues that they were no longer needed. People were upset, and not just the project managers. Several engineers were angry at the change and the way it was handled and Larry was surprised by the pushback they gave him. That reorg was an inflection point at  the company. Afterward, all the engineers (a number that grew into the hundreds) reported directly to Wayne Rosing, the new head of the department.

Larry is a very smart guy. No doubt he learned from this experience that the most expedient course is not always the one that generates the least friction. After years of watching Eric's management, I suspect that he will have mellowed somewhat. Somewhat, but not entirely. For example, the Wall Street Journal ran a story on Larry's recent request to the product managers that they send him an email of no more than 60 words explaining what projects they are working on. A lot of new senior executives would take the occasion of their appointment to meet staff members, chitchat with them to create a personal bond, and inquire casually about their work in progress. Larry approaches things with a more rigorous efficiency.

If Larry were to meet with 100 managers, at 15 minutes each, that would be 25 hours out of his life. If he can scan 100 emails in an hour, he would save himself a full day. Larry values his time more than the need to make his managers feel warm and fuzzy about having face-to-face meetings with him. Unless things have changed since I worked there, most of his managers (at least the ones with engineering backgrounds) understand that completely and would do the same thing. A side benefit is that this puts the onus on the managers to boil down their most essential projects to a few sentences, rather than rambling on about irrelevant subjects. This approach wouldn't work for most organizations, but it's pretty much expected at Google.

Some pundits are asking if Larry's ascension will mean he lets his wild ideas run free, causing Google to lose focus on revenue generation. I have a couple of responses to that. First, Larry has a circle of people who moderate his more extreme visions and keep Google grounded. Chief among this group when I was at the company was Urs Holzle, Google's first VP of engineering and now a Google Fellow. Urs understands both Larry's vision and the constraints of reality and has an uncanny ability to make things happen that would seem to be impossible. But when something truly is impossible, he says so and Larry hears him. 

Salar Kamangar, who is now head of YouTube, was another person who filled the role of sage counsel when I was at Google. Larry respects Salar, who has proven time and again his savvy about business operations and the market for Google's products.  Salar is not a dreamer, but a strategist who sees opportunities that are very much of the here and now. So even though Eric Schmidt might play a lesser moderating role than previously, there is a band of trusted advisors willing to tell Larry that his vision is correct, but the timing is wrong.

The second answer to the concern about Larry's "out there" ideas is that given a choice between a visionary leader and a manager solely focused on improving next quarter's results, most talented engineers in Silicon Valley would not hesitate to choose the former. I was lucky enough to chat with Larry one-to- one about his expectations for Google back in 2002. He laid out far-reaching views that had nothing to do with short-term revenue goals, but raised questions about how Google would anticipate the day sensors and memory became so cheap that individuals would record every moment of their lives. He wondered how Google could become like a better version of the RIAA - not just a mediator of digital music licensing - but a marketplace for fair distribution of all forms of digitized content. I left that meeting with a sense that Larry was thinking far more deeply about the future than I was, and I was convinced he would play a large role in shaping it. I would rather jump on board that bullet train than ride a local that never missed a revenue stop but never arrived at anywhere worthwhile.

Larry has had a lot of wild ideas, like launching a search engine in a field overcrowded with them, offering advertisers a self-service product that put up text ads instantly without review, and scanning every book in the world. All of those ideas seemed outrageous at the time, but Larry didn't care. He knew they were the right thing to do and he made them happen. So, my prediction for Google under Larry's reign? Greater efficiency, faster product launches, bigger, bolder initiatives and a lot more head-scratching among those who don't see as far down the road as Google's CEO - a road that could lead Google to unprecedented global influence or a disastrous derailing. 

Either way, it should be one hell of a ride.

Meet the Press Monday

Looks like I'll be cranking up the hype machine for IFL on Monday, the day Larry takes over (again) as Google's CEO. I did an interview with Reuters print and will be appearing on Bloomberg TV around 9:45 AM (Eastern time). Could someone stop by my mom's house in Florida and program her DVR?