Player Roster

JERRY YANG
CEO and Chief Yahoo
Yang cofounded Yahoo, which he created with David Filo in 1994 while both were electrical engineering PhD candidates at Stanford University. He has been a guiding force behind the company's growth since it incorporated in 1995, but only became CEO this summer.

TERRY S. SEMEL
Chairman
Semel became chairman and CEO in 2001, replacing original CEO Tim Koogle. He ceded the CEO role to Yang this summer under heavy criticism over Yahoo's missteps, particularly in search marketing. But Semel is also the leader who helped right the ship when Yahoo was suffering in the wake of the dot-com bust. Previously, he spent 24 years at Warner Bros., where he rose to chairman and co-CEO.

DAVID FILO
Cofounder and Chief Yahoo
Yahoo's cofounder remains intensely interested in the company's technical choices, often involving himself in hardware and networking decisions. But he shuns the limelight and seems not to covet any title more specific than Chief Yahoo.

SUSAN DECKER President Decker has been the driving force behind Yahoo's recent reorganizations, which were intended partly to unify Yahoo's display and search advertising sales forces. Hired as CFO in 2000, she took charge of Yahoo's advertiser and publisher divisions in December 2006 and was bumped up to president in June. Previously, she was global director of equity research at Donaldson, Lufkin & Jenrette, where she worked for 14 years.

ASH PATEL
Executive Vice President, Platforms
and Infrastructure Division Since starting at Yahoo in 1996, Patel has played a key role in architecting and enhancing MyYahoo, Yahoo Finance, Yahoo Messenger, Yahoo Chat and many other products. He is responsible for Yahoo's global technology investments and platform initiatives, overseeing divisions including Product Platform Engineering, Platform Strategy & Architecture and Advanced Development.

USAMA FAYYAD
Chief Data Officer and Executive Vice President, Research & Strategic Data Solutions Fayyad guides Yahoo's overall data strategy, including privacy policy, prioritizing data investments and managing the company's analytics and data processing infrastructure.

A PhD with degrees in both computer engineering and electrical engineering as well as computer science and mathematics, Fayyad previously put his data mining expertise to work for Microsoft Research, NASA's Jet Propulsion Laboratory and two startups of his own. He joined Yahoo in late 2004 and in 2005 championed the creation of the Yahoo Research organization charged with developing "the new sciences of the Internet."

BRADLEY HOROWITZ
Vice President, Yahoo Product
Strategy Group

Horowitz has become a prominent promoter of innovation within Yahoo, which he joined in 2004. As an entrepreneur coming out of the MIT Media Lab, he founded Virage, maker of a multimedia search engine, and served as an executive of Autonomy after it purchased Virage. He has championed key Yahoo acquisitions such as Flickr and promoted programs to encourage innovation, including Brickhouse, the company's internal new-product incubator, and the opening of the Yahoo platform to outside developers.

[A Brief History of Yahoo]
One Step Forward, Two Steps Back

Born in 1994 as "Jerry's Guide to the World Wide Web," Yahoo started life as a Web directory created by Jerry Yang and David Filo while both were in the electrical engineering PhD program at Stanford University. Yang came up with the concept and Filo made the technology work. Yahoo incorporated in 1995 and went public in 1996, a year after Netscape, the original dot-com darling.

Originally, Yahoo was a hierarchical directory assembled manually by a team of editors; to search on Yahoo was to search that directory rather than the Web itself.

For algorithmic search, whereby software analyzes a Web sampling and determines the content most relevant to a given query, Yahoo partnered with others, starting with AltaVista in 1996 In 1998, Stanford students Sergey Brin and Larry Page asked Filo to consider buying or leasing their Google search engine for use on Yahoo. Filo said their technology was "solid," according to Google's corporate history, but encouraged them to grow the service by starting their own search engine company and show it to him again when it was fully developed and scalable.

In June 2000, Google scored one of its first coups, a deal to provide the search results on the Yahoo site. Just a couple of months later, Google announced the keyword- driven ad program that powered its subsequent growth.

But Yahoo was slow to react, which proved a costly mistake: Advertisers soon realized they could get higher clickthrough and conversion rates with search advertising than with display ads. And at pennies per click, search marketing was more cost-effective.

Google used other approaches to get ahead, too, including grid computing to drive down the costs of massive processing and storage required to index the Web. Google spends five to eight times less per user on storage than Yahoo, which uses more mainstream enterprise storage from vendors such as Network Appliance, according to storage consultant Robin Harris. Cutting storage and processing costs is critical to Yahoo's profitability: "If they can get their infrastructure cost into the ballpark with Google, they still have a very strong base to work from," Harris says.

However, Yahoo chief data officer Usama Fayyad says outsiders don't give the company enough credit for cost control, saying, for instance, its deal with Network Appliance is so favorable "it doesn't make sense to use anything else" for many apps.

Yahoo is also promoting the open source Hadoop project, a clone of Google's distributed computing technologies. (For more on Hadoop, see http://go.baselinemag.com/hadoop.)

Portal vs Anti-Portal

By most accounts, though, Yahoo's bigger mistake was not recognizing the revenue potential of search.

While Yahoo was busy building a comprehensive Web portal, Internet users who just wanted to search learned to bypass Yahoo's clutter for Google's minimalist home page, which started as only a keyword entry blank, search button and logo.

It wasn't until 2004, as Google prepared to go public, that Yahoo launched its own search engine. The technology came from a string of acquisitions, including Inktomi and Overture, which had acquired AltaVista.

Even after its 2003 purchase of Overture, which had essentially invented the search advertising model Google was then perfecting, Yahoo failed to aggressively exploit or improve its search marketing technology until fourth-quarter 2006. That's when it finally delivered its long-awaited Project Panama, the system advertisers use to target ads against a set of keywords and organize them into campaigns, bidding for top placement in an auction-based marketplace similar to Google's AdWords.

Panama Problems

The same year Filo met with Google's cofounders, 1998, also saw the rise of GoTo.com, later renamed Overture Services. GoTo.com's twist on search was that it accepted paid listings, encouraging advertisers to outbid each other to capture the top spot. As academics, Brin and Page had worried that advertising would corrupt the objectivity of search results. But under pressure to make money, Google's leaders decided to accept keyword-driven ads provided they were separate and distinct from objective search results. Google also tweaked the advertising model, making the most clicked-on ads float to the top of the listings.

Yahoo countered by buying Overture (eventually renamed Yahoo Search Marketing). But then what happened? The search marketing business continued to grow and bolster Yahoo's bottom line, but efforts to improve it and better integrate it with the rest of Yahoo were left in limbo, according to former employees.

"The Overture leadership had recognized, even well before the Yahoo purchase, that there was a need to rebuild the platform," says Ken Rudman, who was senior product manager for the Panama Project's Web services interfaces.

One reason the overhaul didn't happen quickly is that Yahoo wanted to make basic changes to Overture's technology first to be sure it would scale. But with Overture operating as a separate business unit in the Los Angeles suburbs, itsprojects didn't get the same priority as the Sunnyvale-based parent company's, Rudman says.

The original plan was to keep Overture separate so it could continue serving other sites that outsourced their search marketing (which at the time included Microsoft's MSN) as well as Yahoo. "So it was deliberate, but I think it was a mistake," says product strategy vp Bradley Horowitz. Planning for the Panama Project started in late 2005, building on the ashes of two previous efforts to overhaul the search marketing, and most development was completed during a well publicized push in 2006. "Yet much of what we wound up doing was very similar to what we originally thought we needed to do," Rudman says. The functional requirements changed very little, although Yahoo helped Overture operate on a larger scale, he says.

Christopher Rhude, another project manager who worked on Panama, says when he joined Overture in September 2004, almost a year after the acquisition, "there still wasn't a lot of knowledge of how we were going to move forward, overall, as part of Yahoo. Yahoo thought they had bought a cash register, but it was grossly more complex than a lot of the systems they managed, despite the scale being similar."

Where the goal should have been to integrate Overture's assets into Yahoo, "there was a certain us-vs.-them mentality and dismissive politics on both sides," Rhude says.

To be fair, the Sunnyvale engineers had reason to doubt Overture's ability to execute, Rhude adds. Earlier, stalled efforts to overhaul the search marketing platform had been organized around three-month internal software release cycles. "That led to a break/fix cycle that was very oppressive," he says, because by the time quality assurance saw the code, so many changes had been made, it was hard to track bugs. With Panama, the search marketing developers shifted to shorter, incremental software, which kept the project moving faster, with less time on "triage," Rhude says.

A former Panama quality assurance team member, who asked that his name be withheld, says the project hewed too far toward a sequential "waterfall" software methodology as opposed to the more agile, incremental development style practiced elsewhere at Yahoo. Worse, scheduling was driven by top management's pronouncements to Wall Street rather than a realistic assessment of the project, he says.

When the schedule slipped by just a few weeks and Yahoo failed to deliver Panama in Q3 2006 as promised, the intense stock market reaction made management redouble the pressure to ram the project through to completion.

In the end, these expats say, they were proud of the quality of the system they created, but they weren't motivated to stick around and make ongoing improvements.

Panama was largely a game of catch-up with Google features, but it also included "a lot of innovation under the hood," Rudman says. "If they've done their work right, over the next year it will become apparent."

The application programming interfaces (APIs) Rudman's team worked on, for example, let advertisers place bulk orders (as opposed to entering ads one at a time into a Web-based user interface). Instead of imitating Google's APIs, the Panama team aimed to make theirs more flexible, he says. "The idea was to organize it in a different way, one that wasn't dictatorial about how to use the system."

"It's really clean and well thought out," says Alan Rimm- Kaufman, president of the Rimm-Kaufman Group, a search marketing firm that runs its own systems that connect with the Yahoo and Google search marketing APIs. During development, Panama engineers were receptive to sugggestions that made the product better, he says.

Yet Rimm-Kaufman still thinks Yahoo is about a year behind Google in search marketing sophistication. And when he sees the results Yahoo can deliver to his clients, "neither the quantity nor the quality of the traffic is there," he says. Yahoo shouldn't give up, he says, but he is concerned about whether the company can attract and retain the people needed to continue to improve the system.

Departing employees from across Yahoo say they're frustrated by its bureaucracy and bewildered by its reorganizations, according to an external recruiter who works with the company. Making recruiting and retention tougher, Yahoo can't offer the stock options and other perks Silicon Valley workers have come to expect in return for their long work hours, the recruiter says.

Rudman says it wasn't hard to decide to leave after the Panama system went live: "Most of what I had to deliver was finished in September [2006], and I left in April. I wound up feeling a little distant. A lot of people I respected were fleeing the ship."

Besides not seeing an immediate payoff in the value of their stock options, they felt let down professionally, he says. "I think after something that big, you need to be ready to tell people, 'Don't wander off, don't lose focus, here's the next challenge we have for you.' I did not feel my team was acknowledged, rewarded—like Yahoo cared about us."

Yahoo's Horowitz says the company recognizes attracting and retaining talent as a major challenge and is working harder to act cohesively, better integrate acquisitions and clear away bureaucracy that interferes with innovation.

"Those people who have left wouldn't have seen what I'm seeing," Horowitz says, because it's just happened since Yang took over as CEO. "It's amazing, the transformation that is happening. It's almost if I'm watching Jerry and David refound Yahoo before my eyes."

As for the former employees who lost faith in the company, it's probably better they "make room for others who see the potential," he says.