site and refine their searches, as well as analyzing each series of clicks and where they end up going, Bing engineers hope to improve the decision aspect of the engine.
Shum says his team is working with Microsoft Research on things like statistical language models, to better understand and interpret user queries, and statistical learning and distributed computing techniques to understand relationships between queries and URLs. “Whoever controls natural language technology will control the future of search,” he says.
As for the culture of Microsoft’s search team and its place in the company, Shum was adamant. “It’s very clear to all of us the importance of the success of Bing. The whole online services division is rallying behind Bing,” he says. “We have significant and ambitious goals for what we want to achieve.”
The most interesting point he made is how Bing differs from other Microsoft products— this, to me, seems to be the key to how it’s being received by customers. “Bing is a service, not only a product. It requires a very different kind of engineering mentality,” Shum says. “24-7, something can be wrong on a large scale. It’s not only serving people in the United States, but globally. As head of engineering, my No. 1 challenge is, this is a service. The mindset is different from other products.”
How so? “It is a data-driven business,” he explains. “Users happily share a lot of data with us on what decisions they make. We have to start with the data, understand the trends and user needs, understand all the tasks. It’s not just, ‘Wouldn’t it be cool if we did the following?’ The amount of data is really amazing. It’s all about experimentation. Our job is never done. Every day there are new things popping up. We have to do experiments every day before we reach any conclusions to look at [a given] feature.” (And of course, this process was under way long before the official Bing launch.)
Asked which aspect gives the Bing team its biggest competitive advantage, Shum didn’t hesitate. “Our willingness to innovate,” he replied. “If you’re the market leader, you want to continue to eat more market share. The Bing disruption is not good for the market leader. We will be even more focused on what users want, not just how to make more money.”
For now, Bing is still well behind Google, but Shum is excited to build on his team’s progress. “We are really pleased that we received a lot of positive feedback. We’re also very realistic. As Steve Ballmer often says, we have to be optimistic and realistic at the same time. It’s the beginning of a long journey.” (At the faculty summit this week, Shum said Ballmer recently congratulated him on 10 years of research at Microsoft’s Beijing lab, and “the next 10 years in search.”)
Lastly, I asked what we should expect to see from Bing in the coming months. Shum was coy. “We will surprise you,” he said.
