Jerry’s Take On What’s Next in Search
With Yahoo! approaching its 10th anniversary, the question I’m hearing a lot lately is “what’s next in the world of search?”
Ten years ago, we were focused on a simple yet vast problem: finding better ways to aggregate and organize information so people can find it. Today, the challenge is different. On the one hand, there’s a lot more information to aggregate and it’s not just more in terms of quantity; there’s a larger variety of content as well — from products and images to news and business information. In addition, we’re pulling content from more sources than ever before.
On the other hand, our user’s expectations have also changed. It’s no longer enough to simply provide a structure for users to find what they want on the Web. Today, people expect to find precisely what they’re looking for exactly as it relates to them. It’s the old example of the “Java” search query. Are you looking for coffee or for the programming language? People want to define what’s relevant to them in their own personal way. They also want to tap into the source of their information at will and they want to manage it all to personally suit their needs.
That’s what is exciting about where we are today. Search as a problem is still far from being solved. The user is in the driver’s seat: they want an experience that is increasingly personal, more relevant, and ties into their task more integrally. Search is just a way to get that integrated experience, but it’s all about what the users want – when they want it, how they want it, and who they want it from.
Jeremy hit on it in a recent blog entry; we have to “make search more relevant and personal.” Those two things are the natural progression for search and they are tightly connected to our concept of seamless integration. Search has to reach a higher bar: it has to enhance the user’s life on a daily basis. Integration of search, community, personalization and content builds the foundation for relevancy in people’s lives.
Because the Net is obviously a bigger part of people’s lives than 10 years ago, we at Yahoo! also have an opportunity to integrate into people’s lives more deeply than before. Yahoo! Local and the beta version of My Yahoo! Search are just two of the examples of how we’re enabling people to manage their search content, search within locations of their choice, and build personal communities online. Users can connect to people with similar interests and they can gather and share search information at will.
Fortunately, we’re also at a time when the technology is helping us plug into people’s lives even more richly. For instance, at this year’s Web 2.0 Conference there was a lot of talk about RSS and wireless technology. This is stuff we only dreamed about ten years ago and its helping redefine what we do with search today. RSS is allowing people to access exactly what they want and wireless is letting us deliver the information wherever you are. People aren’t chained to their PCs anymore and neither is search. Yesterday’s introduction of Yahoo! Search for Mobile is just one example of how technology is propelling search forward. Search is literally in your pocket and at your bus stop. It doesn’t get more integrated than that.
The question to ask now isn’t if or when; it’s “what else.” What else can we do to take search to the next level? What else can we do to make search even more useful and accessible to you?
These are the challenges that will keep us busy for at least another 10 years and we’re getting closer everyday. At Yahoo! it’s our job to stay ahead of consumer needs and expectations and, based on the responses of our users, I believe we’re doing a really good job so far — but it’s still very early. It’s one of the reasons I remain really excited about how we can continue to provide real solutions to people’s problems, and make a difference. While I’m not nearly as technical as I was 10 years ago (I got my hint when David Filo changed the password on me so I can’t touch code anymore), I firmly believe that the technology we are building today makes the future of the Web even more useful, informative, and entertaining. As long as there’s a way to help people find more precise and more relevant information on the Web, you’ll find me in the thick of things searching for it.
Jerry Yang
Chief Yahoo
Hey Jerry,
(That should be coming from Kramer,huh?)
What I have appreciated about this blog so far, is the emphasis on new information as opposed to simply marketing material. (Or should I say, marketing material interspersed with new information).
I was really excited from the title, “What’s Next in Search”. But it wasn’t really about “What’s Next in Search”. I think a better title would have been “The Past Few Weeks In Review”.
I’ll still keep reading though….
Natasha Robinson
Real Estate Logic
putting Logic in Real Estate.
http://www.realestatelogic.net
Hi.
I think the biggest current problem for search engines is that the internal used information about websites are not up-to-date.
Search engines like Orase (http://www.orase.com) are the solution for this problem.
Best regards!
“Java” isn’t the best example of a meaning-poor query, since nearly all of actual searches for “java” are for the programming language (people typically search for coffee as “coffee”).
Better examples for the elevator pitch would be “apple” (the top 10 associated queries include travel, computers, clothing and food) or “Paris” (is the user searching for the celebrity, the city or the hotel in Vegas?)
A new book by Amanda Spink, an information sciences professor at Pitt, provides valuable information about users’ search habits: the average query length is two words and the average number of queries per session is two. These behaviors haven’t changed much since 1997, so training users to query is probably not the answer.
Personalization has some promise, by tracking favorite knowledge domains, but probably requires interface complexity similar to an advanced search, so it ends up being another power tool and not a solution for the bulk of users.
Solving the interface riddle by encouraging some very basic refinement of intent (perhaps by displaying clickable synonyms and/or offering grouped previews of content) seems like the way to go.
Another possibility is tagging search results with small icons indicating that the individual returned pages belong to popular content domains such as travel, health, technology, etc. This would add functionality for users to make a quick scan of results without diving into a lot of text.
SOME CONSTRUCTIVE CRITICISM
Yahoo blew the opportunity for search when search was the future. Now that search is the present, Yahoo is on the bandwagon, riding the latest trends.
The problem with that is that by focusing on present service/revenue models Yahoo is shortchanging unlocking future innovations.
Yahoo has a history of clinging to what works today, at the expense of innovating what will work tomorrow.
With great respect, this blog post is more of the same. I heard the exact same thing two years ago from Tim Cadogan and company in a Silicon Valley conference sponsored by the WebGuild. Two years later, no new songs.
>>I was really excited from the title, “What’s Next in Search”. But it wasn’t really about “What’s Next in Search”. I think a better title would have been “The Past Few Weeks In Review”.
Couldn’t have said it better !!
Please don’t try to sell us a damn thing. I thought ysearchblog was more about technology than marketing. Guess the ails of yahoo run deeper than i thought.. UGH.. UGH
In Jerry Yang’s most recent post to the Yahoo! Search Blog, he says “people expect to find precisely what they’re looking for exactly as it relates to them.”
Great thought. I couldn’t agree more. Then there’s reality.
For instance, when people search for a particular site, do they want to be directed to that site’s competition? Does the user consider that a relevant result?
Clearly the answer is “no.”
Okay, the user will probably tolerate a certain number of irrelevant sponsored results, because they are clearly advertising.
However, when even the unsponsored searches produce this result, Yahoo! Search is just not giving the customer what she wants.
One example of this problem hits very close to home for me. As the founder, owner and webmaster of MouseSavers.com, a site that provides information about discounts on “all things Disney,” I’m frustrated to discover that typing “MouseSavers.com” into Yahoo! Search produces, as its number one unsponsored result, a link to a site that not only has no relationship with my site, but in fact is direct competition with my most important business partner.
(By the way, I’ve looked at the competing site’s code. They don’t appear to be doing anything obvious to “trick” the Yahoo! Search web crawlers. For instance, they aren’t using MouseSavers.com as a keyword, page title or metatag. They don’t link to my site. And they aren’t paying for placement, as far as I can tell — if they are, they’re violating my trademark. So it’s a mystery to me how Yahoo! Search would think there’s any connection between the two sites.)
Further, a Yahoo! Search of my site name does not produce any top-level results that link to my site’s home page! I believe a link to my home page is about result #75.
The competition is getting it right somehow. The same search on Google produces a link to my home page as the #1 result.
I became aware of these problems with Yahoo! Search because one of my regular readers wrote to tell me that she used to navigate to my site by typing “MouseSavers.com” into Yahoo! Search (she uses Yahoo! as her home page). She can’t do that anymore, and it drives her nuts.
It seems Yahoo! Search thinks my reader wants “personalized” results that don’t actually link to the home page of the site she searched for!
I pay to have MouseSavers.com listed in the Yahoo! Directory. I’ve reported these bad results several times. But it doesn’t get fixed.
Jerry Yang asks in his blog, “What else can we do to make search even more useful and accessible to you?” How about starting with a search result that actually links to the site people are trying to find?
Mary Waring
MouseSavers.com
Charlene Li of Forrester wrote
“Overall, I believe that the true value of Yahoo! Local and also, of Yahoo! Web Search on mobile devices is tie-in back to Yahoo! The instances when you would actually need this type of information will be far and few between, but when you need it, lo and behold, Yahoo!s there to provide it for you.”
The real money is still in search and the face of search is spreading to many fronts, in the way soda pop spread from the druggist’s counter to the vending machine and mini-fridge in myriad places.
The recent experience with my personal blogs has made me much more conscious of the search engine space. Until now, to me, Google has been synonymous with “search,” and MSN and Yahoo! were non-entities. That changed when I started reading ysearchblog, poring over my http server logs and doing searches on my writing. I came to appreciate the maturity, competitiveness, and hunger of these alternative search offerings. However, my referrer logs show that Google still clearly dominates the space.
One of the “tie-ins back to Yahoo!” created by Yahoo! Local and search on mobile devices is the blooming in the public mindset of the meme “Yahoo! means search too.” In this frame, these two offerings are mindset conversion tools.
I’ve seen media buys in conventional media for Yahoo! Local. Presumably they are aimed at everybody, from current Yahoo! service users to non-users alike. Contrast that with Google’s reliance on free PR and capital investment and extension of search into the blogging, email, newsgroups, and alliances.
If Google had Yahoo’s portfolio, you can be sure they would be extending search into Yahoo Groups, Yahoo Mail, Yahoo Finance forums, and Yahoo chat. I haven’t been to Yahoo chat in ages, so I can’t report on what is going on there, but the first three are still rich untapped conversion opportunities.
The existing search function on Yahoo Groups and the Yahoo Finance forums basically suck. The search only seems to cover the few recent posts and is good for one keyword. The rich trove of information and discussion on those servers is largely inaccessible; compare that to groups.google.com. (I’d have to pull the posts down from Yahoo with a perl script and have Google Desktop index them. BTW, check out Google’s Group 2 Beta, groups-beta.google.com.)
Each of these services represents a relationship with a customer who also keeps a position map, ala Al Ries and Jack Trout, in their head. So when I think stocks and stock information, I think finance.yahoo.com but the shitty search there makes me think Yahoo! = !search, and when I do have to search for something, I naturally don’t think to use the !search people. That is the position battle Yahoo has to fight en mass and that is why pouring marketing dollars into Yahoo Local! is likely to be ineffective in challenging Google’s dominance in my referrer logs.
Jerry,
Yahoo! is so strong with user profiling I figure you’re ahead of the game in knowing how to profile a search. This is great for the user but there seems to be a huge gap between your user profiles and market related results. When do you see marketers catching up with Yahoo and what can they do now to better optimize for personalized search in the next ten years?
Thanks,
Brian Oblivion
http://www.helvis.com
American Punk Gone Wrong
Jerry,
A couple of questions for you.
a) Why isnt yahoo directories being pushed-considering it addresses the searcher concerns of relevancy, inspite of tedious efforts to get at it?
b) On search results–how does yahoo search work w.r.t recency of results ?
Personalisation? I’m wondering if this area is being a little over-thought.
How about working on providing a more powerful tool so that the users can extract information more easily themselves, rather than trying to second-guess them?
For example, cluster the results and prompt the user with clarifying questions. “Did you mean java as in coffee or java as in programming language”.
So if Google’s not building a browser then why are they appearing on my sitemeter report?
http://thomashawk.com/2004/11/if-googles-not-creating-browser-then.html
In Jerry Yang’s most recent post to the Yahoo! Search Blog, he says “people expect to find precisely what they’re looking for exactly as it relates to them.” Some of the points in Jerry’s post and the subsequent comments seem to be pointing towards behavoral analysis in order to deliver relevancy.
This IMHO is a fundamentally flawed argument as it will result in the views of a minority imposing their views on the majority.
What is needed is an understanding of the language being used or in short a linguistics approach that produces an element as yet unknown in the field of
search: context. With a sense of understanding the context of a query, results can be filtered to be entirely relevant to the query. The ‘Java’
example is a good one and it is worth bearing in mind that with the growth in domain name sales, virtually every word in the English language has more than one meaning now.
This is a new paradigm for search but will certainly be the way forward.
Have you seen the work done by linguist Prof. David Crystal and his company Crystal Semantics? They have some great linguistics based products for search and seem to deliver context solutions which solves problems like the many meanings of ‘apple’. Worth a look.
Personalization can be overused both semantically and operationally.
If a system is always in training to “personalize” results based on past history then by design is will also narrow – exclude – results based on PAST precedents – not necessarily what is currently being researched / purchased / or wanting to know about NOW.
No different than the newspaper editor’s way of selecting stories. The wire service makes a decision as to what it will put onto the wire, the editor then eliminates what he thinks people are not interested in, then elimiates what there is no room for, and only after that do users get to eliminate what is left – but this is after three layers of filtering has already been done that most are not aware of.
The way personalization is being thought about and implemented now is effectively doing the same thing.
As long as users have a way of turning it OFF so as to take on the responsibilty of being their own editor will I think personalization – filtering – will be useful.
Of course, in dealing with masses of people, most people do not want to take on that responsibilty and are happy to let others dictate what is filtered out and what is ultimately delivered to them.
Tom Philo
http://www.taphilo.com
Congratulations on your excellent work in recent weeks. We had sort of given up on Yahoo because almost all of our traffic was coming from Google. We are now seeing that things are changing, at least on our website/blog. We are seeing that the results produced by Yahoo search are much more accurate now and we get high quality traffic.
Keep up the good work. Competition is good for all.
Is there anybody at Yahoo! to talk to about the search function of Yahoo! Groups? For a major search engine like Yahoo! to have such a poor search utility is out of character with the rest of the site. Is this a bug, oversight, what??
Many thanks in advance,
-Bob
Is it true that that name “Yahoo!” does indeed come not only from acronym for “Yet Another Hierarchical Officious Oracle” but from the classic book Gulliver’s Travels by Jonathon Swift as well?
The question to ask now isn’t if or when; it’s “what else.”
This is a very good observation. “What else”, The truth of the matter is that we will never be able to fix these problems in the context of building webpages and websites. The web is much to primitive to solve problems of complexity. Something is missing. That “What else” is a technology called V2. You wont have to go to the internet via the web, it will come to you via V2. Check it out at http://www.Webkiller.net
I was intrigued by the topic as the rest here … What’s next ? I would like to ask a question – is there a possibility of 3D becoming popular and people searching on shapes & 3D information in future? How long will the market take to adopt this concept?
This is interesting site, got good information about site. Even I have similar site as would give few information about my site…..Develop and implement strategic and tactical web site focused marketing strategies that will accomplish your Goal…..If you want more information about site click the link marketing