Yahoo and Mobile Search

rbaeza-mayo-2006.jpgOne of my fellow speakers at the Barcelona chapter of Mobile Monday was Ricardo Baeza-Yates Director of Yahoo! Research Barcelona. Ricardo had to leave early, so we didn’t have much of a chance for a chat, but there were some very interesting aspects of his presentation.

I think it’s fair to say that Ricardo¬†sees mobile as an adjunct to the web, as opposed to a completely separate channel and therefore his presentation was about search generally, with a few choice snippets about mobile.

The main principle about how Yahoo! approaches search is the “Wisdom of Crowds”, which means search results are honed based on tags, links and previous results to search queries. This means that they’re doing an awful lot of work into how items get tagged. In other words, if they have a picture of a banana and I’m looking for “fruit”, I’m not going to find it.

One really interesting game that explores this issue is their ESP Game. You play with an anonymous and uncontactable collaborator, by simultaneously being shown an image. You both type in tags until you have a match and then move to another. You have 2 1/2 minutes to get through and agree as many images as you can. It’s not very easy actually and demonstrates some of the complexities of accurate or intuitive tagging, as well as being quite fun.

When Ricardo moved on to talk about the difference between search on a mobile and search on a PC, there were a few surprising differences. Search on mobile uses more words, despite the comparative difficulty of inputting. I’d guess that this is down to the user wanting to be as accurate as possible, first time, rather than subsequently wanting to refine things.

The other idiosyncrasy of mobile search is that there’s more variety in the queries submitted. I can’t remember the exact figure (does anyone know?) but it was something like for PC search, the top 20 queries account for about 50% of searches, whereas in mobile,¬†the top 20 are¬†only 2%.

All the big online guns (Yahoo!, Google and Microsoft) are focusing their sights on search as the key to to unlock mobile revenues and it may well be important. I have a feeling that it won’t be quite as important as everyone assumes.

My theory is that what people want on a phone is discovery, rather than search, at least at this stage in the game. So a¬†mash-up of “Jerry’s Guide to the World Wide Web for Your Mobile” (ironically),¬†an element of serendipity, a dose of peer ratings and smidgeon of your friends’ recommendations. This isn’t a million miles away from the Wisdom of Crowds that Ricardo spoke about in terms of how results are delivered, but maybe it calls for a different user experience.

[tags] ricardo baeza-yates, barcelona, mobile monday, yahoo!, search, discovery, mobile [/tags]

—–>Follow us on Twitter too: @russellbuckley and @caaarlo

Share this post:
  • Facebook
  • Twitter
  • del.icio.us
  • StumbleUpon
  • Digg
  • LinkedIn
  • Google Bookmarks
  • Suggest to Techmeme via Twitter
  • Reddit
  • Perrin Samuels
    Surely there's some truth in the claim that "what people want on a phone is discovery," but it's hard to say how much because "discovery on the phone" is incredibly ambiguous. What IS discovery on the phone? It's more like REsearch than plain old search.

    What I'm looking for is answers. It's hard to imagine this isn't the case for others. When I think about getting answers to arbitrary questions, at arbitrary times, about arbitrary subjects, I have to think hard... Where do I think I can find the answer I'm looking for? Better yet, who do I know who I think is available now, would know the answer, or would be able to find out quickly, or point me in the right direction? In many cases, I fully expect to have to do some work to get a decent answer. I might have to speak with several people, read some stuff, travel, and possibly experiment a bit. And most importantly, I'm completely willing to accept that I may have to apply some judgement of my own to formulate an answer that is satisfyable to me (and maybe nobody else) from pieces and parts of information that, separately, leave me with more questions than answers.

    That's the tricky part! Sometimes--quite often!--the responses I get, and information I find, during early phases of research, generates more questions than answers.

    Getting good answers and explanations often means formulating them yourself. It is as much an art as it is a science. It is, indeed, discovery, and discovery takes a lot of work. It is often not possible to offload this work. There's a philosophical issue at stake here. Only the discoverer may discover; everyone else is removed by more or less understanding, depending on what they already know. I'm exaggerating of course, but the point is this: Search is done. Let's stop talking about search and start defining the new problem space. The next big challenge is to solve the problem of discovery--ie, REsearch--in a collaborative way, and by some degree of automation. Now that would be hot!
  • Hi Russell,

    Hope you are doing well.

    I think one key difference between the Web and Mobile search may be the location from where the search is being performed. For example, if I am searching for an item while I am in a mall I am interested in finding the closest stores where the item is carried. I guess this is not an easy case to implement, but it would be great if the mobile searches can get there in a few years.
  • I wrote a post several months ago on 4INFO's blog talking about the challenges of mobile search in a world where people have already adapted to web search. It is a challenge of both improving our own service based on analysis of user queries and teaching the user how best to locate "focused" information from their cell phone.

    http://4info.typepad.com/blog/...

    Your thoughts on people wanting "discovery" is interesting since I am not sure that we are quite there yet. My take is that people want immediate answers. Not that it will always be this way. As we adapt to our phone's capabilities, then "discovery" will likely become the norm. I will no longer carry books and magazines on my travels, since I will be able to access them from my phone, perhaps. At that point, 4INFO not only becomes the avenue for answers, but a segue to discovery when we combine both focused information and related news.

    Not only must you consider the search, but also the result. In a recent CNN Money article, our Amol Joshi was quoted as saying, "People want answers, not links. Google has the disadvantage of being a Web search firm."

    http://4info.typepad.com/blog/...
blog comments powered by Disqus