A few days ago there was a Techcrunch interview of Google’s Eric Schmidt. Here’s the bit that struck me:
[TC] The long term goal of Google search, he says, is to give the user one exactly right answer to a query:
[Schmidt] So I don’t know how to characterize the next 10 years except to say that we’ll get to the point – the long-term goal is to be able to give you one answer, which is exactly the right answer over time.
The one answer? Are you kidding me? That’s Google’s long term goal? That’s the extent of their imagination when it comes to information seeking? Granted, even that narrow problem isn’t solved yet, so I’m not against more work being done in that area. But where the love for Exploratory Search? Search is so many things, including but not limited to learning, comparing and contrasting, synthesizing, discovery, planning and forecasting, as so on.
All of these information needs go beyond finding the one answer. Sometimes there is no one answer, and the goal of the search is the discovery that multiple answers exist. Sometimes you don’t even know what question to ask, and the goal of the search is the accretion of enough knowledge so as to be able to ask the right questions in the first place. (See Belkin’s ASK model of information seeking, in which it is explicitly acknowledged that a single text is likely not sufficient for satisfying a user’s anomalous state of knowledge.)
Yet Google’s long term goals do not include supporting anything other than “the one right answer” finding? Unbelievable.
Daniel Tunkelang read this same interview and sees a bit more hope than I do. He sees some of Schmidt’s comments as suggestive of an increased willingness to engage in HCIR, which itself is more exploratory by nature. I don’t see it, though. The CEO of the corporation has spoken clearly: “The long-term goal is to be able to give you one answer.”
That sounds no different from what Google is trying to do right now. The long-term goal does not include any growth into or acknowledgment of other forms of user information seeking behavior. Rather, it is more of the same of what they already do now. Just piled higher and deeper.
Sigh.
I totally agree with your take on Schmidit’s comments. However, he says “..I don’t know how to characterize the next 10 years except to say that we’ll get to the point.. “, lets hope that by that time he will realize that “giving one right answer” is definitely not the way to go.
The pithy comment is most likely a distillation of Google’s roadmap and could mean a thousand things. An engineer many levels below from Schmidt could probably extrapolate the comment into a book of a thousand pages. It all makes for good PR.
Secondly, is there really one right answer to a query? For example, say I pop over to the local Apple Store to use one of their machines because it doesn’t know me and enter the query ‘sunshine’ into Google. What is the right answer at that moment in time?
@dinesh: First, your point about it being a distillation comment is well taken. However, I think there are dozens of better ways he could have said something equally pithy without expressing the exact sentiment that was expressed.
For example, instead of saying “The long term goal is to be able to give you one answer”, Schmidt could have just as pithily said, “The long term goal is to satisfy your information need”. Or “your need for information”.
Satisfaction of that need could then come in a myriad of forms: One right answer, a set of answers, a set of better questions, a characterization of a space, a comparison and a contract between competing answers, etc. If your information need is to have a comparison, then so be it. If your information need is to have one answer, then so be it.
But you can be pithy without sacrificing accuracy. How does it make for worse PR, to be both pithy and accurate, as opposed to just pithy? Unless Google’s goals really do not include doing anything but “one answer” search.
Second: When you ask whether or not there is one right answer, are you talking about results personalization? The right answer for me is not necessarily the right answer for you? Sure, I believe that. But that still doesn’t change what Schmidt said. He said, “The long term goal is to be able to give you one answer”. The “you” includes, in my understanding, results personalization. But the phrase as a whole still does not include exploratory information needs.
@Raza: Yes, I hope that they do realize that there is more to information retrieval than the one right answer.
I was making a general point (independent of Schmidt’s PR) about entering a query and the search engine returning the right answer. If I go into my local Apple Store (or for that matter any publicly available machine that I’ve never used before) and enter the query ‘sunshine’ what is the right answer that the search engine is going to send? Sure, with personalization the search engine is likely to return the right answer but that is not new news. I’m parsing the absurd here ie. can a search engine return the right answer for you without knowing something about you beforehand?