Interview Exclusive: Bing Search’s Stefan Weitz
Stefan Weitz, the Director of Bing Search, thinks people should expect more out of a search engine. I do too.
When Microsoft CEO Steve Ballmer unveiled Bing back in 2009, some of the world expected artificial intelligence (AI) to spring from their PC monitors, another segment was skeptical, still more people reacted much as they probably would if someone unveiled yet another search mechanism today – with total indifference. But there was a time when web search was all a buzz with potential 100 or more search companies vying for what was then, and is now, Google’s crowning web achievement. Powerset, hakia, Search Wikia, Mahalo, Yahoo!, Lycos, Cha Cha, dozens and dozens of others, and not one has come close to surpassing Google.
Stefan Weitz is what some would call a “geek’s geek”, and it’s his job to make sure Microsoft’s Bing does get close. Weitz mission is to direct Bing Search development to the next level of search technology – to evangelize Bing, and to gather the bits and bytes of user feedback for ostensibly improving user experience – and vastly. This is the round about way of saying Weitz is Bing’s biggest fan’atic’, and he’s darned proud of it. A 12 year veteran at Microsoft, Weitz in fact seems like a man with a mission, not only to improve Bing, but to find the perfect gadget period. Remember, I said he was a super geek.
The Q & A below reflects how comfortable super geeks converse on their favorite subjects.
Q – Having tested and reported extensively on natural language search engine Powerset, via PowerLabs back in 2007, I have to start by saying, compared with the expectations we all had of semantic search, Powerset, hakia, and some others, Bing did not exactly get off to a stellar start. My question is Stefan, have slow adaptation and too great expectations hurt Bing’s success so far?
A – As you know from your reporting days, semantic search is both technically difficult and seldom used by users. Most queries still hover around the 2.6 word mark as people have become accustomed to engines giving them terrible results when they ask more natural language questions. That said, it’s fascinating to see younger people like my 8 year old daughter interacting with search technology. To her natural language is…well…natural. As more people engage with search on more devices without keyboards, the notion of keyword-based search as the ultimate manifestation of the billions of dollars spent on the tech will become outmoded.
To that end we have invested heavily in building out the understanding of the natural world. We think of the web as a digital proxy of the physical world – not just a collection of pages and links. Our Satori system launched last year has been quietly building a model of the planet and has hundreds of millions of entities and billions of facts and connections. Finally, we are working within the constraints of how users today interact with search technology. In other words we asked ourselves if people are still only using a couple of words to search, how can we do a better job understanding what they are really asking?
The technology that powers our intent derivation engine is pretty stellar (if I do say so myself) and allows us to use over a thousand possible signals to really understand what you’re asking for when you search “ORD to SEA” and present you with an answer that can help you do something, not just see more links.
Q – How difficult has it been for you to overcome what we might call “negative stigma” in gaining audience?
A – It’s not necessarily a negative stigma that we have to overcome – it’s peoples’ habits. When we launched Bing we asked people how consciously they thought about their choice of engine. The act ranked between tapping their leg when bored in a meeting and deciding on whether to brush their teeth at night. In other words, the act of searching and choice of where one searched had become autonomic – people generally just didn’t think about it.
So our challenge is really to break through those habits – to get people to demand more from search than presenting a bunch of links in response to a keyword. It’s why we’re investing so much in multimodal experiences where Bing simply becomes part of the fabric of your day whether it’s on your television, within your productivity suite, on your mobile device, or on your tablet. We think the act of search should weave itself into the fabric of your daily experiences – not be something you ‘go do’.
Q – You probably hear way too much of this, but I know a great many people who would like to know your view on Google’s continued dominance of search. In particular, “do you guys talk in the board room about this, especially given the semi-collective efforts of Bing, Yahoo!, even Facebook?” I know people might be fascinated to understand your thoughts.
A – As I mentioned above, Google is a habit. Yes, they built a great product and continue to introduce innovation in the space, but it is very much built on a model that was designed for the web of yesterday: one of links and pages. They can easily continue to have a great business in that space. Their challenge is, that when you have a great hammer, everything looks like a nail. Their experiences on mobile devices, tablets, even TVs all look like keywords and links, and that makes sense given that 97% of their revenue comes from ads they serve along side of those links.
Bottom line – we recognize we have to approach search differently. We understand people are using search to do more than get information – they are using it to do things in real life. They are using search as their broker between the physical and virtual worlds and that demands fundamentally new ways of combining the different graphs that exist on the web today. We need to utilize the social graph as people rely on other people to make decisions. We need to use the geo-spatial graph to understand the real world in which we live, and to recommend the correct actions. We need to understand the application graph to connect people with services that can help them take action when and where they need to. And, we need to wrap all those things up in an interface that makes sense given the context of the user.
Q – Bing has a great many striking features millions of people like a lot. For me the aesthetics of Bing, even the initial organizational capacity of Powerset itself, were and are appealing. You guys have, in my view, literally turned over heaven and Earth to enhance search. Can you talk about your favorite features of Bing?
A – Thanks! It’s a really fun space. My favorite features would have to be the ones that do a great job pre-processing the web and helping make sense of things – not just dumping more information on users. Some really cool examples that don’t get enough attention are things like product reviews inside Bing Shopping, where we literally comb the web for thousands of reviews on products, semantically analyze the reviews, and break them down into their component parts. The net result? If you care about picture quality of your TV most of all, we will show you the comments – both positive and negative – about that aspect of the TV. It means we save users the time of reading hundreds of reviews to see if what you care about is mentioned somewhere in all those reviews. We literally find the needles in the haystacks and present them to you.
I also love our new Snapshot functionality that identifies entities on a search results page (like Honda Accord) and blends information from across the web about that entity to help you make decisions. Even cooler is when we’re able to identify actions one can take (like making reservations at Wild Ginger) to help you complete your task right from the results page. And last, I really love the work we’re doing in the social graph space. Only Bing is able to combine the digital footprints of your networks so you’re able to tap into their knowledge when you’re searching for nearly anything.
Q – Really agnostic search users, the few there are, would profess I think that Bing is “at least” as relevant and useful at finding things on the web as Google is. Some I know even say, more so. Again, we’re talking “switching” here, but my question is about a “tipping point” – the place where Google users migrate to Bing. Is this something people at Microsoft or Bing discuss? Can you elaborate at all?
A – You’re right. Our Bing it On challenge helps people see what millions of others have told us in independent surveys – that people prefer Bing’s results by 2:1 over Google. The bottom line is that we have to earn people’s searches. And we’ve already done it to an extent as we’ve doubled share since we’ve launched. The next frontier is what I mentioned earlier: weaving search into the fabric of the every day. Our vision is that ‘Googling it’ or “Binging it’ will sounds as outmoded as people saying they are ‘going to go online’. You will simply rely on search to be there, in context, whenever and wherever you need it and it will respond to you in a way that matches your intent.
Q – Since we are really discussing conversions Stefan, can you talk about what sort of adaptation numbers Bing is getting off the Opera and Mozilla browser agreements? How about the original Facebook collaboration, how effective market share wise has that been really?
A – We don’t have specific numbers to share but I can tell you it has been great to see the browser communities embracing Bing as a choice in their products. For Facebook we power their web search and of course we have our exclusive deal that lets us tap into the social graph to power social search experiences that don’t require users to sign up for yet another social network.
Q – One of the SEO experts close to us suggests Bing actually now has the best opportunity ever to play catch up with Google, in particular because of “perceived” mistakes made with Google Search. Can you comment on this moment in time, is Bing ready to seize such an opportunity?
A – As you know, this technology area changes more rapidly than almost any other. Every day there are new data types, new use cases, and new devices that constantly challenge the status quo of how search is done. Companies that rely on one model for search will lose share not because they are bad, but because the web is changing more rapidly than they can adapt their business model. At Bing we can take some big risks with existing business models, user experiences and partnerships since we are focused on the long game in search: making it ubiquitous and universally useful. We know if we can deliver that, the revenue will follow.
Q – Is “Bing it On” as a result of such a strategic opportunity?
A – As I mentioned, we need to help people break the Google habit…. and that’s what it is these days. It isn’t that they are inherently better than us in search. Certainly they beat us sometimes, just as we beat them sometimes, but overall our core quality is as good or better than what they offer, and we have things like social data and an investment in task completion that they simply don’t have. Bing It On was all about getting people to try Bing and see for themselves – without the Google brand halo – what we have to offer. Do we win every-time Nope. I have friends who’ve mailed me their Google Win results telling me to work harder. But you know what? Even if we lose, we still win because people think better of our offering than they did before Bing It On – so we love it.
Q – Bing Webmaster Guru, Duance Forrester, just announced “Bing Webmaster Guidelines”, supposedly to help people understand Bing search marketing from a very broad perspective. Some are criticizing Bing for not providing a “rich” enough webmaster experience. How do you address such criticisms?
A – Duane is a machine. He is always on forums and Twitter to address questions from the community. And he takes feedback better than anyone else I know. I do think we are way more open than our competition about how we operate and where we’re headed. I think the new Guidelines are really good. Can we do even better? Sure. But there is always the tension about being prescriptive versus giving guidance. That and realistically what we say today – especially as we build in more adaptivity into the search results based on intent and user behaviors – might not even be wholly useful tomorrow. Give us feedback – we don’t hide from it.
Q – We’ve covered a bit of ground here Stefan, if I may, “where do you see Bing’s market share by the Summer?”
A – That’s partially up to you and your readers! I’d ask everyone to come and try us again. Challenge us, tell us where we are not doing as well as you expect and let us know what you like. We’ll keep pushing the envelope in areas that aren’t traditionally measured by agencies and we’re ok with that. I can tell you we do have some pretty exciting new work up our sleeves by Summer. :)
I’ve done quite a number of interviews over the years, it’s always fascinating to speak with decision makers at the top of their game. The old saying; “there’s something to be learned from everyone”, surely applies to super sharp people as well as next door neighbors. I asked Stefan some fairly tough questions compared with a those of some other interviewees, and I was pleasantly surprised with how forthcoming he was on all points. Or it could be, questions we think are really complex or tough, really have simple and honest answers. Either way, I learned a lot from Stefan Weitz. I mean this.
When I first asked Stefan for this interview via Facebook I thought to myself; “He’ll say no, he’s to busy and important.” Stefan came back within a few hours with a friendly “glad to”. When Stefan did not finish the questions or answer straight away to my reminders I thought again “He did not like the questions, he’ll forget like some others, the politics of search will get in the way, his PR said BEWARE!, whatever.” Stefan was in South America, Tierra del Fuego or Machu Picchu for all I now. When he swore in his next mail to catch up with me in 24 hours, I thought “An old familiar, yeah right.” Before the 24 hours was up, he came through. When I learned Weitz’s favorite saying from his Grandfather (one my Grand Dad had too), I was no longer surprised at his candor. The saying goes:
“That boy isn’t worth the powder it’d take to blow him clear to hell.”
And now here’s the thing, if I believed everything I ever read about Bing’s Director, if I believed all my past perceptions about search were still valid, if Bing were still a fumbling Powerset infant needing some PR mouthpiece to prop it up – well, I would be really wrong about a lot of things. Old habits are hard to break, old perceptions die hard, switching browsers or toothpaste is sometimes painful – but sometimes necessary. Maybe it’s time for a new search, one worth the powder and lead?