Microsoft Moves on from Yahoo Again; Talks Internal Search Innovation

After the latest round of attempted negotiations with Yahoo and investor Carl Icahn, Microsoft is once again saying that it is moving on from trying to acquire the second-place search engine.

Addressing the attendees at Microsoft's financial analyst meeting, Ballmer explained Microsoft's Plan B for building search and competing with Google. And like Joel McHale on E!'s The Soup, I've read the transcript so you don't have to. Here are the sound bites.

First up, the reason why Microsoft thinks it can compete in a Google-dominated search marketplace:

Search is ripe for innovation. It has not been the most innovative category in the world. I mean, think back, what did search look like five years ago, 10 blue links on the left, some ads on the right, and maybe some ads on the top. What does it look like now, 10 blue links on the left, some ads on the right, and maybe something on the top. It is ripe for innovation. If you say to yourself, five years from now, 10 years from now will search be as humdrum, hard, 50 percent of searches don't actually lead to an answer to somebody's problem, is this an area that's ripe for innovation, in user experience, natural language, semantic understanding, consumer experience? The business model hasn't been touched.
Give Google credit, they invented the business model that supports the modern search business, and yet it hasn't been touched. How do you involve the consumer? How do you move to a pay-for-action model? How do you reward the consumer, and involve the consumer economically? This is a category that's ripe for innovation. And that's important, because if it's not ripe for innovation, we shouldn't be doing what we're doing. We will not be able to be very successful by only doing what the market leader does.

Next Ballmer says that Microsoft is ready to get in the ring with Google:

Second strategy for us I call it focus, but in my own mind I think of it as kind of our Mohammed Ali, float like a butterfly, sting like a bee strategy. Pick focused areas of search, really innovate, change those areas, differentiate from the market leader. That doesn't mean Google won't come back and blah, blah, blah, blah, blah.

Ballmer says blah a lot.

And last, but not least, we will work to reinvent the user model and the business model. I know this from past experience, it's often harder for the guy who is market leader to reinvent their category than it is for somebody who is, so to speak, the Avis of the business, number two and trying harder.

Of course, competing in search is all about making a profit from ad dollar$.

We need more relevant ads. It turns out that one of the big advantages the market leader has is they have more advertisers in their system.

You don't say?

So if you look at two pages today, one from Google and one from Microsoft, the thing that's perhaps most interesting is, because they have more people bidding on advertising, they have more opportunity to serve up a relevant ad. A lot of our discussion around Yahoo! really centered as much on this issue as any other issue. How do we get enough advertisers to have a pool of advertising to change the whole advertising approach.

Buy AOL, whose Platform-A is the leading online ad network? Pony up the money for Yahoo?

And last, but not least, we're going to have to invest in search and advertising, in fantastic brand, and fantastic marketing.

That's two last but not leasts. Not sure which not least is the most not least.

Later, Ballmer handed over the stage to Satya Nadella, Senior Vice President heading up Microsoft's search, MSN and ad platform engineering efforts.


When it comes to best results, we have to ante up to create the core infrastructure. That allows us to have a very rich index, a deep and broad index, that helps us actually experiment on the relevance improvements, which is a continuous game. You can think of it as, like, drug discovery, where you continuously are in the game of improving your relevance.
So we have made progress on all of those. In fact, if you look at it, last fall is when we reached a huge milestone for us. We increased the index many-fold. Our relevance improved. We today believe we are comparable to the best in the United States when it comes to the core relevance. And that was not something that we could have said last year at this time when you were here. So that's substantial progress.

Nadella echoed a Yahoo researcher's recent article on Semantic search. Microsoft acquired semantic search company Powerset earlier this year.


We think the Search experience is going to evolve to really understand the query intent, as well as the document content, in rich ways. So this is where we're building the algorithm and capabilities to be able to get to that deeper level of semantic understanding, natural language understanding, so that we can really help users of search engines with their tasks, because nobody does queries in isolation. Queries are always done in the context of some task.

Microsoft has big plans for image search:

One of the nice things about our image search is that we've been able to sort of take an experience where we have infinite scroll capability, so you can – you don't have to paginate to be able to see, you know, hundreds of thousands of images. You can just keep going down. And we have a nice experience for infinite scroll here.

Want to refine those images searches? Ok.

We have the ability for you to refine by size. So if I'm looking for a specific size, I can refine that and find that. So in the case I'm looking for some large images, I can look for color. So if I'm looking only for black and white, I can go ahead and hit black and white. And if the demigods smile, I will actually get that. So, in any case, the idea is to be able to have rich filtering capabilities with your image search to get filtering.
The other thing that I can also do is to be able to drag and drop. I can open up the scratch pad and I can go ahead and drag and drop these images and make a new collection, because if I wanted to come back to a collection of images that I used in this project, then I can have a series of collections.

No offense, but that's not innovative. That's done by Google and Flock, too.

That's ok. Let's just move on to video search.

One of the things that we have done, working with MSR, Microsoft Research, is to do what we've done, all of us in the Search business have done for Web search, in terms of creating previews of content before you click on the link. We've done that for video. So we crawl all of the videos on the Web.
We created something called the Smart Motion Thumbnails, which, in fact, are basically summaries of video content out there. And we call them Smart Motions because we can actually do scene detection. So if it's a news clip, we can actually summarize the news clip. If it's a video of a sports event, we can make sure that we do the scene detection to get the complete shot, and what have you. So this is – so all you have to do is hover over the video. You get the summary before you actually click on it. And this is, again, best in class or first in class when it comes to creating video summaries that are semantically smart.
Just to compare that to what Google has, Google has, you know, a video search product where they have images. They don't have the summary capability. The one other difference is the selection. If you look at our selection, we have from MSN, AP, MSNBC, you know, and CNBC, ESPN -- lots of different sources, as well as YouTube – whereas the selection in video for Google is a lot more biased toward their own owned-and-operated indexes.

Nadella moves on to talk about travel search. This must be part of Ballmer's ideas about focusing on specific parts of search and innovating them:


Whenever you're booking a flight, one of the specific things you're looking for is to time your purchase. We know that airline fares are volatile, so the idea of being able to purchase when we think is the best time to get the best fare is fairly critical. So in this case, if I go ahead and click flights from Seattle to San Francisco, what I do is I get back on something called an Instant Answer. It says that there are many results of flights from Seattle to San Francisco. It also has this prediction algorithm.

Enter in the Farecast acquisition and re-launch.

In this case, it looks like it's a good time to buy. Let's go ahead and look at the actual flights. It takes me to something called a Smart Calendar. I can actually pick the date pair, so when I want to leave and when I want to come back. So let's say I want to leave Thursday and I want to come back on Saturday; I can go ahead and do that. I will hit Find Flights. And at this point, Farecast is going out to all the agencies and all the airlines and bringing back search results that meet the criteria of the dates I picked.

But Microsoft's plans for the HP Toolbar is where they hope to take some precious market share away from Google.

How do we really get more people to know about Live Search and get the taste for some of the value, like, in particular, the cashback value? So the place where we are innovating is in the toolbar. We have recently done a distribution deal for our toolbar with Hewlett-Packard. So this is the toolbar that Hewlett-Packard will carry with some customizations of their own. It's the MSN toolbar.
And so let's say I'm on Google and I type in Xbox. I can go ahead and search for Xbox, and automatically the toolbar detects that you're searching for Xbox on Google and a Gleam view that there is a cashback on Live Search. And so I can go ahead and at this point click on that Gleam and it'll take you to Live Search, or it's supposed to take you to Live Search. Oh, it is on Live Search. See, I didn't even notice the transfer. So it's so seamless that now you're on Live Search. You can get the cashback for a particular Xbox that you want to buy. So that's just an experiment on how we get the word out, get more users trying Live Search, and getting the value of things like Live Search cashback.

What do you think of Microsoft plans? Can they gain market share? Let it fly in the comments.