Yahoo! Scientists Explore the “Three Dimensions of Search”
Ever wondered if searches for topics such as “stock quotes” are made mostly by people from a high income bracket? Or whether it’s only young people who search for “miley cyrus”? Web search engines serve millions of people worldwide – yet, we know very little about who is searching for what, and how they are searching. In a study, recently presented at the WSDM conference in Hong Kong, two Yahoo! research scientists paint a detailed picture of the data behind web search.
Their study addresses the three dimensions of search: “who” is searching, “what” they are searching for, and “how” they are choosing to search.
- The “who” dimension uses demographic properties, such as gender, age, income and education level to obtain anonymous, aggregate estimates, based on user-provided registration data, and by crossing user-provided zip codes with census data.
- For the “what” people are searching for, queries are mapped to Yahoo! directory topics and for each user a distribution across the various topics, such as Recreation/Sports or Entertainment/Music is obtained.
- Features such as session length or number of queries per session are considered to give insights into “how” people use web search engines.
One key problem that is addressed in the research is that results in current search engines are typically optimized on a per-query basis, without taking into consideration who issued the query. So, in a sense, web search engines attempt to know everything about queries while knowing only very little about users. The analysis is only two dimensional. Looking at how the three dimensions (who, what, how) interact therefore becomes a focal part of this study.
The insights presented in the study are interesting from a sociological point of view and, ultimately help Yahoo! deliver more relevant, personalized content to its users.
Some of the aggregate data used in the study is accessible via Yahoo! Clues.
To view the full study by Yahoo! scientists Alejandro Jaimes and Ingmar Weber and, please click here.
- 19 Comments
- Subscribe
Hello I certainly agree with these things
Now people only search for short word or search by product word for specific what they want to know. Search engine never stop for learning what trend for searching.
Yes,It’s very fruitful and easy criteria of search.I like it.
Amazing idea,I like it because it access the quick desired queries with no wastage of time.
thanks for sharing the information i am agree with you
This gonna make search site work differentlly
How do you get this user information? Do you inform your users that you are monitoring and recording their searchs?
The “who” part of the equation, and how that information is gathered, is interesting. What effect is that data having on the results?
This is amazing. Hope Yahoo can further develop this new tecnology.
You’ve got ‘Who’, ‘What’ and ‘How’, but what about ‘When’?
For how – it would be good to analyse how many searches under using mobile devices and how many using standard computers.
A lot more people are sharing information using social media sites and using mobile devices.
It’s difficult to know ‘who’ is doing the searches if many people are using the same computer.
It’s important to analyse who is searching what and when, so that you know who your target audience is.
There’s a lot of money in information
You know the old saying … knowledge is power!
as helios said knowledge is power thank you for giving me power
until now the social network is the number one for sharing information
this is old news , search engines are way ahead now