About me: I worked at Microsoft Bing back in 2008 and I’ve seen it at the peak of it’s power. I was working on a team reporting to Gaurav Sareen, and my roommate was indirectly reporting to Rohit Wad – the two founders of Bing. I moved to the valley about a 18 months ago, and tried a long list of failed ideas. I’m now the co-founder of a fast growing viral startup — likealittle.com.
This story is about an amazing group of technologists, who were on a task of solving the hardest problem in the world — attacking Google at its home territory. Let me say that again: attacking a market incumbent on an area of its core strength – It’s hardest in technology companies. Time and again people have tried doing it and failed, and it’s not just about the money. Quoting Paul Graham:
The “next Google” is unlikely to be a search engine, however, just as the ‘next Microsoft’ was not a desktop software company. I used competing directly with Google as an example of a problem with maximum difficulty, not maximum payoff. Maximum payoff is more likely to come from making Google irrelevant than from replacing it. How exactly? I have no more than vague ideas about that. I wouldn’t expect to be able to figure out the right answer, just as I wouldn’t have expected anyone to figure out in 1990 what would make Microsoft irrelevant.
As another comparison of the difficulty, Microsoft exerted so much market power to overthrow Netscape, and they detected and fought it from the very beginning. Even then, it took “Bill Gates” himself to focus on it (and still shell considerable amounts of resources). They totally didn’t see Google as a threat, until it had a huge market cap. (Don’t be evil, was a joke?) Long story short: It was an informal project inside of Microsoft started by Rohit Wad, who recruited Gaurav Sareen and started writing the first pieces of code for MSN search along with Michael Burrows. You heard that right, Michael fucking Burrows.
They hired some of the finest technologists in the world, and really shook the ground. Bing was growing faster than expected, beats market forecasts and after several years of work, the quality was getting closer and closer to Google — They were narrowing it down endlessly. Then, the “grown-ups” came in, brought in process and discipline, and everyone who knew something, left. Not surprising is it? It’s not a new story. This shit happens all the time.
I’m writing this to appreciate the breakthroughs that the early Bing team made (and more so about Gaurav, since he hired me and I worked in his team of 40). They were informal and they broke rules. Microsoft has specific directions for teams in the US to not hire in other countries competing with the local branch offices. Gaurav, went to India and hired some of the best hackers right out of school, ignoring Microsoft’s directions.
Back then, nobody really offered positions in the US right out of Indian schools. This was totally new in the market. Microsoft also had limits on the amounts you can pay for a particular position. Bing was extremely lavish in compensation, making offers to the best hackers for $90K/year when the adjacent teams were making $75K/year offers. This was the time when Google was making $80K/year offers. They were hiring folks as if everyone in the team was promoted right when they are hired, which was a crazy bet: that economics would still make sense.
Microsoft’s DNA is: slow release cycles and building “solid” products. To the contrary, Gaurav is notorious for saying, “you ship or you suck,” taking the emphasis away from the framework builders. Your promotion was solely dependent on the amount of stuff you ship and the impact you make (that’s not too much news to Web startups?.) They had weekly release cycles — faster than Google back then. The rule on the ground was if code breaks or does not ship in time, the engineer is held responsible, No — not the test, not the project manager. Engineer should drive everything.
They had awfully broken infrastructure. Build systems can take hours. They did not fix it, not because they didn’t know how to. They spent their time shipping code to end users instead, because that’s what would matter in your performance review. This was all phenomenal. The laws of the Web working at a gigantic company, as big as Microsoft. WTF.
These guys were extremely hungry to succeed, and were really smart. I guess, that’s what happens when you pick a young guy down the chain, who is smart and give him a low budget to work on something he loves. He will go out there and with his passion, convince the smartest guys he knows to join him. They will be young and unproven because the budget can’t pay a legend.
The key ingredients: smart guys, love for the product, hunger to succeed, triggers some kind of an explosive growth. For a long time the early teams will hate the word, work-life balance. When the rest of Microsoft is sleeping, Bing-ers were checking in code. Seven years later, both Gaurav and Rohit would be made “Partners” at Microsoft, one of the fastest career growth trajectories in Microsoft corp’s history. The business is making waves and expectations are built. Steve Ballmer declares search is his top priority.
As Ballmer’s ambitions build up, he makes some of the big corporate re-orgs. Harry Schum who was head of Microsoft Research, China would take over Bing as its new leader. They decide to spend $100M in marketing and rebrand live.com to Bing.com, and purchase powerset for $100M to bring in some of the finest search architects world has ever seen.
There was an inorganic infusion of people at the top, including Chad Walters, etc. And then something somewhere went wrong. Things were not working, and I’m not knowledgeable enough to understand why. (My best guesses follow.) When you hire people from outside and assign them top ranks, it takes time for them to gain respect within the team. Meanwhile, the odds of the hire “failing” pose a significant risk to the team morale. Also the rich and accomplished come to Microsoft to settle down, and lack the “hunger” (in Steve Job’s words).
There was interference to bring in “professional managers,” a part of whom were non-technical. There was push from the top to add more project managers and testers and it was hard to any longer isolate the blame to a failure. Meetings became longer and more parties got involved. Evenings looked emptier than before. Oops.
In technology, it takes very little time to go irrelevant. My manager has moved to Zynga. Everyone whom I rated “good” in my team have left Microsoft now. The best ones whom I’m in touch with are looking forward to leave. Hell, even Rohit and Michael have left to Google. Gaurav is still in the team and well respected, but he works on extremely different stuff now. I have a feeling he would be the next one to flee, although he is very passionate about the empire he created. Who knows, he might sign up to be the last man, standing on the sinking ship.
As for the quality, the team’s offers have been “standardized” and they pay far below market rates now. Companies like Facebook and Zynga have opened Seattle offices and will scavenge the remaining good parts of Bing. They are in no race to compete in hiring quality anyways.