We’re at OutCast Communication’s offices for a Yahoo Search event that they’ve dubbed “The End of the 10 Blue Links.” It looks to be a state of the union for Yahoo’s search product, and a look ahead.
There’s a clear theme already in this presentation: Search is shifting away from links to intent, according to Yahoo. And it’s moving from just pages to objects. They don’t just want to serve up a series of static results based on a word, but rather want to leverage the data they’re collecting from products like Search Monkey and mobile search to figure out what people are actually looking for.
None of this is hardly a new idea, but Yahoo feels it can do the best job of it as it’s been working on many of these technologies for a while now. And with products like BOSS gaining big usage (over 30 million queries a day now), there’s a lot of data coming in.
Below find my live notes:
Prabhakar Raghavan, Head of Yahoo! Labs and Yahoo! Search Strategy
- It’s time to kill the 10 blue links.
- Take the insight from mobile search and take it back to the web.
- Figure out the underlying intent of the user.
- We need to move from a web of pages to a web of objects.
- It goes from how many pages we index to how complete of a picture do we make.
- We need to build a web of objects from a web of pages.
- We need to use the wisdom of the crowds — it’s not just about machine algorithms, but the people who gives us information through Search Monkey.
- BOSS is another important part — open up the search engine structure for other players to use it.
- Query volume of BOSS been growing fast, past 30 million now.
- Microsoft’s offering is doing 40 million queries a day, BOSS is hovering around 30 million already
Larry Cornett, Vice President of Consumer Products
- Understanding user intent, creating open search ecosystem and mobile
- Let’s talk about intent: Users have goals that go way beyond a search page: 1) Intent around keywords 2) Intent around objects 3) Intent within time
- In July 2007 we kicked off some of our new initiatives (search suggestions), launched way before competitors followed suit
- The drop-down menu when you do a query, this does query-completion and explore related concepts.
- Rolled this concept out to image searching as well. But image searches are done visually, so we include pictures in there.
- All of this is available through the BOSS API
- Currently in testing: Search for “Paris” and get images of Eiffel Tower. Bring in objects from the real world.
- Also in testing: Recognize an object in the real-world and figure out what they’re looking for — again, all about intent.
- The final dimension: Time. Search Pad is the product we’ve been using to do research. People don’t recognize that they’re doing the same query every night for 6 months. Search Pad is being bucket tested.
- There is intent being expressed over time. “iPod” is different from “iPod fix”
- Get music to play right on page, and music videos to play right on search page — old stuff, Yahoo just patting itself on the back.
- 1st anniversary of Search Monkey.
- In 23 markets around the world. 70 million enhanced results viewed everyday
- 15% increase in click-through rates for some partners
- 15,000 developers using it
- 400 applications in the gallery
- Increase in structured data 413% since we launched
- It’s completely open-standards compliant
Alex Moskalyuk, software engineer for Facebook — Yahoo Search Monkey partner
- Building a Search Monkey app was a valid project for us
- People want to see a user’s picture and their bio information — we can also display the geo information
- Facebook had to create a special URL schema for profiles to work with Search Monkey
- They also had to format pages in hCard microformat
Matthew Hertz from pipl.com — BOSS partner
- Pipl started in 2004, public beta since early 2007 — People search engine – we covered them here.
- We’re good at deep web, but we needed surface web data — that’s where BOSS came in
- Also use BOSS for spam detection and algorithm screening
Marc Davis, chief scientist Yahoo Mobile
- 4 billion people have a mobile device around the world – we’re looking at a world that people use mobile to meet fundamental needs
- Another Star Trek example, it needs to be real easy to figure out what you’re looking for especially on mobile.
- We’re talking to a web of objects that can be connected to by people
- Answers, not just links — with a minimal amount of clicks as possible
- Understanding user context — these 4 billion phones know a lot about you already, now we just need to connect that with results
- Showing off oneSearch — oneSearch shortcut, etc. Search assist is a “lifesaver” on mobile.
- 1 billion PC users versus 4 billion mobile users worldwide — not all Internet users yet, but will be soon
- In Indonesia, mobile search volume is about 4 times greater than PC search. Emerging market growth is juge
- 70 partnerships in 40 countries.
- This is not about one phone (*cough* iPhone *cough*)
Summary and What’s Next?
- Ending the 10 blue link legacy to move to user intent
- Creating experiences from objects and their relationship with the world (rather than just pages)
- Leading the industry with an open foundation for innovation
Q&A Session
- Interfaces in testing — when will we see them? We learn from these bucket tests — if they test well they can launch in the coming months.
- How does a web of objects change the way you serve ads? It doesn’t really change it, but web of objects can help with intent which helps with advertisers.
- Why is this different form what competitors are doing? Eventually users don’t want to read documents, so we want to surface the web objects. A non-answer
- When would it be appropriate to show a page or a link in this object world? Every users won’t want the same thing, so we need to better tailor to cert
- Real time question — the sexy thing? — A bit of a buzz word. Need to separate that word from just Twitter. There’s much more to real-time. Product plans: there is definitely an information need fulfilled by real-time, but they’re saying just something right now. Don’t sacrifice relevance. We’re investing in being able to discover fresh content as fast as possible. Find the freshest content we can and show it. Blend the best of both
- Yahoo Answers, how to use that in search? We should use both and figure out what people are looking for.
- Microsoft new search product with these qualities? For us, if something isn’t testing well, we’re not going to launch it, not based on what others are doing.
- Privacy? It’s opt-in, we don’t have to save the results.
- BOSS model has changed right? Open monetization was launched earlier this year.
- How accurate is intent? Search assist helps with that.
- Tech behind bucket testing things? Our technology plus information we gather from Search Monkey, etc.
- What constitutes an object? Anything in the real-world.
- How do you group objects together? We don’t insist on things being tagged, but there is a lot of data out there to tie things together. And there is value in curation.
- Is it fair to call this semantic search? Yes, that’s fair — at least somewhat. But this goes far beyond static pages and documents. Connect the web of documents to the web of the world.
- Moving from BlackBerry to iPhone? We’re on a thousand different devices globally — we love the mobile web. On a BlackBerry I use Yahoo oneSearch the client. We’re focusing on where we see consumers going — so in going where consumers go.
Yahoo search = automatic fail.
Much better than Google and Microsoft’s in terms of APIs. (I don’t even think Live search has an API)
Live Search has a very robust API:
http://blogs.ms...-per-month.aspx
And MG, where is Yahoo getting those qpd numbers? According to comScore, Live search serves 2.6B queries a month.
10 blue links are gud for novice net users.
Couldn’t agree more.
You see so much innovation coming from Yahoo and the quality of their results is very good. Given all this, only business issues could stop them from succeeding.
it’s good to see yahoo innovating again
So there is Yahoo trying to come back in the search business… well some ideas sound interesting, but lets see.
These points seem incredibly vague – I’m not sure we can really extrapolate anything of consequence from them. The search engines have always talked about focusing on user intent more, I’m not sure how this is any different – this seems more like marketing lingo than technical.
they are fairly vague, just some notes now. i’ll interpret in a bit.
I’ll look forward to it. Wasn’t criticizing your notes, more so just being skeptical of the claims being made by Y!
understood for sure. i feel the exact same way.
What exactly ARE the ten blue links?
hyperlinks
is that an answer? doesn’t answer anything to me. I’d also like to know what do they mean by ten blue links.
The 10 links on the first page of the Search Engine Results Page, I assume…
that’s a good answer. thanks
Ah, an answer at last!
Thank you Nick and Miguel, that does make sense and seems a bit obvious now
Actually, I just gave the SearchMonkey-based add-ons a try on Yahoo, and I have to say … I’m impressed. Just reset my default search engine back to Yahoo in (ironically enough) my browser of choice, Google Chrome.
So, come on. If we’re really going to have knee-jerk brand reactions to stuff, we deserve the dark, backwards, depressing Internet free of new ideas that we’ll get.
“Search is shifting away from links to intent” is still an idea? Does Yahoo implement any intent-oriented search already in its search engine? How is its algorithm looking like? Or, is it just a publicity strategy? I bet Yahoo guys are implying that search is heading toward semantic search, like the algorithm that powers Powerset.
Search is trendy again! Wehoo, love it.
Ok, so, can they please move past search submit pro? I’m sick & tired of this program, more than likely it violates FCC guidelines for ads…and the number of tech people I talk to that have *zero* idea Yahoo still does this is apalling.
When I worked there, I campaigned incessantly to shut it down. Hasn’t happened yet.
Unless & until they nuke that program, Yahoo Search is firmly embedded in the web of 7 years ago, when search engines charged publishers to scrape their content. That’s putting the cart before the horse like nothing else.
Does anybody in Silicon Valley *use* Yahoo Search? Most people I worked with at Yahoo (granted, I left in 2006, and stopped consulting with Yahoo in 2007) none of my coworkers used Yahoo, they all used Google.
“Figure out the underlying intent of the user. ” yahoo chief – search strategy.
the mystery is overwhelming. dudes probably still thinking how yahope needs to know the difference between paris hilton and the effil tower. or how pizza means a picture of one or how to order one. knowing what a consumer wants is not rocket science. Strategic keyphrase “natural language location” need i say more. i sincerely hope this guy is not serious.
PizzaLocator.com – order now
no, ahole, you do not need to say more. You need to say less — much less. You are an idiot and we all hate you.
blowhardlocator.com — I can’t believe you were allowed to live.
your whistle is weak.
BornLocator.com – cant help it
assclownlocator.com — look in your mirror
What is the BOSS model?
Well, it all raises an interesting question… yes, looking at the end-user, default Yahoo search, there’s not a whole lot there that isn’t there on Google — and there needs to be a LOT there for people to switch.
But as those numbers imply, Yahoo can be a better search tool *for developers*. For instance, if I’m building new site search tools and want to build upon an existing API, I may not particularly care if the back end comes from Yahoo or Google. I just want whatever works best. And that could mean that even if Yahoo isn’t stealing search volume from Google.com and the Google box people set to their browser default, they *could* wind up adding up volume on other sites. And as a developer/publisher, I’d really rather give people enough functionality that they don’t immediately go to Google.
But I agree with other comments here — let’s see more of this in search itself. I still see those ten blue links on Yahoo, which is part of the reason no one is bothering to leave Google. And both Google and Yahoo (and Live) win if there’s some real competition and forward motion. Aside from some subtle improvements, the whole set of tools has been painfully stagnant, and I don’t think that benefits anyone – even Google.
And Yahoo shouldn’t simply be dismissed for following Google – Yahoo started before Google as a directory in fact
If people want the 10 blue links they are going to Google. That game has been won. So I don’t think it is such a bad idea for Yahoo to take a chance and go in a different direction. I think they should try to display search results in a way that’s similar to Kosmix, but better. It sounds like that’s what they are aiming for. Hopefully they will be bold about it and give us something new and different. It can’t be all Google all the time.
Try the qwiji Search result experience: http://www.qwiji.com
Search the Web | Videos | Images | Web-Shows , instead of filtering search result, you just flip through the actual most relevant pages. Enjoy.
So should we call this s the beginning of search 3.0?
Google = new features coming out
Yahoo = new features coming out
Wolfram = new
MSN = ?
SEO gimmicks just won’t cut it anymore…. you are going to need real content.
Another Yahoo day session of feature speak and not enough focus on what consumers actually need. Sounds like the same big Internet co. story. Newsflash: Consumers don’t have any burning need in Search. When they fail, they keep trying. Why? Cause it’s so easy and fast.
I’d suggest Yahoo spend the thousands of dollars it takes to have an all-day session, event, whatever you call it on understanding what tweens are doing and what their needs may be now and in 5 years. There are smart people at Yahoo, and I bet they could figure it out.
I believe any move towards providing better search results, especially when they take into consideration the “reason” or “because” factors behind the search, will tend to please web surfers. Two other priorities, in my opinion, should also be addressed. Deep web research and anonymous surfing. See my blog today where both are featured. http://Surfin4u.blogspot
Comments regarding improved web search results including references to “reason” or “because” factors are my opinion as author of a blog at
http://Surfin4u.blogspot.com thank you.
Yahoo’s goal is interesting. The notes here, however, do not say anything concrete.
Bing’s take is similar—going beyond simple web pages as result. Including specialized search results for different query intentions.
We are quite excited to see the increasing “verticalization” of search. At Cazoodle, we specialize in developing vertical search engines for specific domains—e.g., Apartment Rentals, Online Shopping, Local Events, etc.