I just noticed this post by Sarah Perez on ReadWriteWeb: “Google: “We’re Not Doing a Good Job with Structured Data“. Here is the excerpt that caught my attention:
The company wants to be able to separate exploratory queries (e.g., “Vietnam travel”) from ones where a user is in search of a particular fact (“Vietnam population”). The former query should deliver information about visa requirements, weather and tour packages, etc.
While I’m not sure how important it is to automatically distinguish between “exploratory queries” and fact-finding ones, I absolutely agree that neither Google nor its web search rivals have delivered much in the way of exploratory search capabilities to users.
Alon Halevy, who is leading up Google’s “deep web” efforts, seems to understand the problem, even if I might quibble on some of the details. The question is: how does he plan to solve it? And how will his solution integrate with Google’s decidedly non-exploratory approach to information seeking?
8 replies on “Is Google Serious about Exploratory Search?”
I remember talking to one of Google’ s very knowledgable search guys in November of 2006. I told him I was working on exploratory search. He had no idea what that was/what that meant.
It’s a big company, and I know that most Googlers are systems hackers who aren’t all that knowledgeable about information retrieval, let alone exploratory search. But they also have Dan Russell: http://dmrussell.googlepages.com/
I suspect that, given Google’s size and ability to attract talent, they have at least one of everything. The question isn’t what they can do in theory, but what they will do in practice.
No, but this guy wasn’t a system hacker. He was someone who, before joining Google, had published at SIGIR. Repeatedly.
To me, that indicates that, once inside Google, even non-system hackers become system hackers, and lose perspective on what’s going on outside the Google “IR = known item search” bubble.
I think “Exploratory Search” was not, back in 2006, a common usage term even in the IR field. But I think anyone having a minimum IR knowledge given a quick definition of that concept will have said “oh yes I know that.”
Anyways, for sure Google is experiencing things around those concepts. They already answer factual questions differently. You could also do product search, music search -> music:”artist name”, etc. etc.
Because Google main driving directions are still governed by the minimalistic approach, they insert things very carefully, remove them, retry, etc. Trying to understand users’ intentions is not an easy thing and going too far in that directions leads to queries misinterpretations and in the day-to-day Google business it’s probably not helpful.
Remember that keyword search is working “great” for Google in a very high percentage of user queries even for the exploratory ones. Google prefer let the user do their explorations by offering them appropriate site(s) to do explorations. (try query vietnam travel in Google for ex.) When appropriate, expert sites on the exploratory subject are tough to beat. Not perfect but as far as I’m concerned still the best in most cases for day-to-day searches.
P.S. Don’t get me wrong, I don’t think Google minimalistic approach is the magic recipe to all IR problems. Semantic search, guided navigation, browsing, more complex search UI, vertical search (ex.: travel web sites UpTake)… are still needed and there is large spaces where Google day-to-day sauce is simply just not good (ex.: Google Search Appliance)
I agree that it’s not fair to put too much stock in a vocabulary test. Indeed, I suspect that most IR folks still haven’t given much thought to exploratory search, by any name.
I also acknowledge that Google is doing quite well with its minimal model, perhaps in part because no one has mounted a serious challenge to it. Yahoo and Microsoft are striving for parity, which is a lost cause.
I do question their emphasis on understanding users’ intentions vs. communicating with users. Perhaps Google’s Achilles’ heel is the hubris that makes it so reluctant to ask what it thinks it can guess–like the stereotype of a man unwilling to ask for directions.
My day-to-day search tool is not Google. Rather, I use the esoteric and unfortunately named “Duck Duck Go” engine (http://duckduckgo.com/). It’s not perfect, but it’s the first credible exploratory search engine I’ve seen for the web. But of course they have no chance of making a dent in the market. I just hope they can stay alive (and that Yahoo! BOSS doesn’t pull the plug) so I can continue using them!
Finally, as you point out, what is nice to have for the web (which at least has Wikipedia) is a must-have in the enterprise. Not everyone has figured that out yet, but it’s clear that people are getting there.
DuckDuckGo looks pretty interesting. Will try it.
[…] do note that, other than the deep web research, which I covered in an earlier post, I don’t see much about how Google is innovating in search. Perhaps Google is done with […]
[…] blogger, named Daniel Tunkelang, recently brought up a discussion, within a discussion, on determining the exploratory nature of a query. He questions whether this is worth it. This is a very interesting focus. Ryen White has published […]