Q&A: Windows Live Search Senior PM Justin Osmer

On Monday, Microsoft decided the fundamental functionality of its Windows Live Search platform was complete, and rolled out the 1.0 version of its all-new search engine as the replacement for MSN Search. MSN has been, over the past several quarters, Microsoft's least profitable division, but now the company's goal is to put on a par with those companies that make Internet search one of the most lucrative businesses in information technology today.

There will be significant handicaps to overcome. The very name "Windows Live Search" seems to proclaim it's not for the Linux or Mac user, even though its services are presented in industry-standard JavaScript (AJAX). If Windows Live is to be a revenue source for Microsoft, it has to be recognized by advertisers as a viable, active evolving platform.

This means it has to acquire market share, and with Google's share now estimated at 60.2%, #2 Yahoo's share at 22.5% and Microsoft splitting the rest with "Other," according to Internet media analyst firm Hitwise, Microsoft has a long way to go. Advertisers track performance in terms of overall reach, so the terms "distant third" can be a serious turn-off.

But does Windows Live Search really have to be #1 at some future date, before anyone declares it successful? Can't Microsoft find itself in a more comfortable position with, say, a close third? BetaNews sat down with the company's senior product manager for Live.com, Justin Osmer, to find out, and discuss a new service called Windows Live QnA.

BetaNews: When I've talked with analysts about Windows Live Search, and what they interpret to be the goals for Microsoft, they try to frame it in terms of how fast or how soon Live Search will be able to "beat Google." I'm thinking in my mind, it doesn't have to beat Google to be successful, does it? Couldn't Live Search be comfortable as a #3 player?

Justin Osmer: Certainly, third place is not a bad place to be. Given how far we've come in such a relatively short period of time, especially compared to our competitors, in building our own home-grown technology, we're pretty happy with third place. But we also are not going to rest on those laurels, because there is more money to be had, there's more advertising dollars available out there, and with more market share comes more advertisers.

It's a bit of a cycle, and we absolutely want to continue to grow our user base, continue to grow our market share, and that certainly is a goal behind this recent release - to bring to the market some new offerings as part of the search experience that are different. That will help people get to the information that they're looking for in a unique and more customized way.

BetaNews: Your strategy, I believe, appears to be to position yourself as a tool provider, much the same way Ask.com has done fairly successfully since it dropped the "Jeeves" from its name.

Justin Osmer: We certainly believe that search doesn't have to be a one-way experience, so we're trying to put in more controls in the hands of the user, to allow them to get that direct access. A great example of this is with our image search capability. We've spent a lot of time during our recent beta period listening to customer feedback. So, some of the things that we've done with image search that we believe is kind of best-in-class, is to remove the metadata from the search results that are shown on the screen.

What we found is people wanted to see images, they didn't want to see a bunch of text links next to the images. So we allow you to resize those images in a very fluid way -- the page doesn't have to reload, it's all based on AJAX -- then if you want the metadata, all you have to do is hover over the image and then you get the metadata - the size of the image, what Web site it's from, that sort of thing.

Then if you like that image, you can simply click and drag it down to what we're calling Scratchpad, which can be a collection of your results so you can go back and quickly find the images that were of most interest to you. It's a version of an editor's table at a photo desk. It puts them all in one place for you to review them, so you don't have to worry about trying to find them back through the search results.

BN: Whenever Google introduces a new feature to its search page, that feature generally leverages its core, back-end power, its colossal index. They've described it to me before as an "organic growth scenario." Has Microsoft learned from that somehow, and in repatterning MSN Search as Windows Live Search, is it going to be integrating an "organic growth scenario" in the future?

JO: Certainly, you learn a lot from looking at the query data, and seeing the types of queries that people are entering, and when they have to go back in and enter a new query. One of the main objectives we have is to stop abandonment, where people will go in and get a query, maybe do another query, and they'll leave frustrated not having found what they were looking for. That's something that we're continually working on.

Today, we have a feature called Instant Answers, which pulls data from Encarta, as well as Fox Sports and a number of other internal and third-party Web sites. It offers up factual information: "What's the population of Washington, D.C.?" "Who won the Super Bowl in 2002?" Things like that pop up as an instant answer at the top of your search results.

Certainly, as we continue down this path, to use the term, the organic growth opportunity there is just to leverage those different elements you have across those different categories of your index, so they're all coming up in the same place. The Instant Answers scenario will start to prove itself out with News. If you go to Web and do a news topical related search, you may see a news result at the top from our News vertical, but then you'll also get the general, inline Web results.

BN: You talked about wanting to stop the abandonment, to avert people leaving without finding what it is they're looking for. Sometimes, when you have a search engine that is, well, too good at what it does, and it comes up with accurate results on the first try, then you have a scenario where customers are satisfied and then leave, as opposed to other customers who are dissatisfied and keep searching and keep searching.

JO: Exactly, that's one of the reasons why we do a lot of lab testing with actual users, and we're watching them click through the service and supplement that with the raw data as well. Because you're exactly right - Instant Answers is a great example. If I want to know who won the Super Bowl in 2002, I just type that in and get my answer. I probably won't click on anything on the page, I'll move on. That was a successful search result for me. Being able to compare that to someone who types in a query, and looks at the page dumbfounded, can't figure out what to do, and leaves - those are the types of things that we're constantly working on.