Google Shared. A Case-Study in Social Search

TLDR: A novel and practical approach to Web search that combines ideas from personalization and social networking to provide a more collaborative search experience is described, which offers considerable business potential in a Google-dominated search marketplace.

Abstract: Web search is the dominant form of information access and everyday millions of searches are handled by mainstream search engines, but users still struggle to find what they are looking for, and there is much room for improvement. In this paper we describe a novel and practical approach to Web search that combines ideas from personalization and social networking to provide a more collaborative search experience. We described how this has been delivered by complementing, rather than competing with, mainstream search engines, which offers considerable business potential in a Google-dominated search marketplace.

Figures

Fig. 4. The HeyStaks search portal provide direct access to staks and past searches.

Fig. 7. (a) Average staks created and joined per user. (b) The percentage of sociable and solitary users.

Fig. 5. (a) Average staks created and joined per user. (b) The percentage of sociable and solitary users.

Fig. 1. The HeyStaks system architecture and outline recommendation model.

Fig. 2. Selecting a new active stak.

Fig. 6. A representation of the collaboration network among HeyStaks searchers.

Full text

Provided by the author(s) and University College Dublin Library in accordance with publisher
policies. Please cite the published version when available.
Title Google shared. A case study in social search
Authors(s) Briggs, Peter; Coyle, Maurice; O'Mahony, Michael P.
Publication date 2009-07-10T09:25:48Z
Publication information Houben, G.-J. ..[et al.] (eds.). User modeling, adaptation and personalization : 17th
International Conference, UMAP 2009 : Proceedings
Conference details UMAP - 1st and 17th International Conference on User Modeling, Adaptation, and
Personalization, Trento, 22-26 June 2009
Publisher Springer
Link to online version http://dx.doi.org/10.1007/978-3-642-02247-0_27
Item record/more information http://hdl.handle.net/10197/1244
Publisher's version (DOI) 10.1007/978-3-642-02247-0_27
Downloaded 2022-08-09T20:30:12Z
The UCD community has made this article openly available. Please share how this access
benefits you. Your story matters! (@ucd_oa)
© Some rights reserved. For more information, please see the item record link above.

Google Shared.
A Case-Study in Social Search
⋆
Barry Smyth, Peter Briggs, Maurice Coyle, and Michael O’Mahony
CLARITY: Centre for Sensor Web Technologies
School of Computer Science and Informatics
University College Dublin, Ireland.
{firstname.lastname}@ucd.ie
Abstract. Web search is the dominant form of information access and
everyday millions of searches are handled by mainstream search engines,
but users still struggle to find what they are looking for, and there is much
room for improvement. In this paper we describe a novel and practical
approach to Web search that combines ideas from personalization and
social networking to provide a more collaborative search experience. We
described how this has been delivered by complementing, rather than
competing with, mainstream search engines, which offers con siderable
business potential in a Google-dominated search marketplace.
1 Introduction
For all the success of mainstream Web search engines, users still struggle to
find the right information quickly. Poor search productivity is largely a result
of vague or ambiguous queries [6, 8, 20], and there is co ns iderable research on
different ways to improve result selection and ranking. For example, researchers
have looked at ways to bias search towards special types of information (e.g.,
people, research pa pers, etc.); see for e.g. [9]. Others have attempted to profile
the preference s of searchers in order to deliver more personalized result-rankings
[10, 11, 21]. Recently, other researchers have explored how to take advantage
of the collabo rative nature of search [1, 12–14, 17]. In our own research we have
explored a collaborative approach to personalized Web search [4 ,18,19], profiling
the preferences of communities of users, rather than individuals, a nd generating
recommendations inline with community preferences; see a lso [7].
While results have been promising, little attention has been paid to the issue
of deployment and it is difficult to see how these technologies can be successfully
brought to mainstream search. We have previous ly explored different deploy-
ment options [2, 5 ] as a way to loosely integrate community-based search with
mainstream se arch engines. However it has b e en clear for some time that nei-
ther approach is likely to work for consumer Web search: users want to search
as normal using their favourite search eng ine. However, the recent arrival of
⋆
This work is supported by Science Found ation Ireland under grant 07/CE/I1147.

browser plugins has presented a new opportunity to deliver third-party search
technology, via the browser, on top of some underlying service like Google.
This paper describes how this has been achieved through a new commercial
venture called HeyStaks (www.heystaks.com). HeyStaks places a n emphasis on
the potential for collabora tion within Web search as a route to a better search
exp erience; see also [1, 12–14, 17]. The key motivating insight is that there are
impo rtant features missing from mainstream search engines. For example, recent
studies highlight that for 30% of searches the searcher is loo king for something
that they have previously found, yet search engines like Google offer no practical
support to help users re-find information. Similarly, for up to 70% of searches
the searcher is looking for something that ha s recently been found by a friend
or colleague [19]. And, once again, search engines like Google offer no support
for the sharing of search results. Helping s earchers to organise and share their
search experiences could deliver significant improve ments in overall search pro -
ductivity. We describe how HeyStaks adds these missing collaboration features
to mainstream search engines and present results from a recent usage analysis
based on the initial beta deployment of the system.
2 HeyStaks
HeyStaks adds two basic features to any mainstream search engine. First, it al-
lows users to create search staks, as a type of folder for their search experiences
at search time. Staks can be sha red with others so that their searches will also
be added to the sta k. Second, HeyStaks uses staks to generate recommendations
that are added to the underlying se arch results that come from the mainstream
search engine. These recommendations ar e results that stak members have pr e -
viously found to be relevant for similar queries and help the searcher to discover
results that friends or colleagues have found interesting, results that may other-
wise be buried deep within Google’s default result-list.
As per Fig. 1, HeyStaks takes the form of two basic components: a client-
side browser t oolbar and a back-end server. The toolbar allows users to create
and share staks and provides a range of ancillary services, such as the ability
to tag or vote for pa ges. The toolbar also capture s search result click-thrus
and manages the integration of HeyStaks recommendations with the default
result-list. The back-end server manages the individual stak indexes (indexing
individual pages against query/tag terms and pos itive/negative votes), the stak
database (stak titles, members, descriptions, status, etc.), the HeyStaks social
networking service and, of course, the recommendation engine. In the following
sections we will briefly outline the basic operation of HeyStaks and then focus
on some of the detail behind the recommendation e ngine.
2.1 System Overview
Consider the following example. Steve, Bill and some friends were planning a
European vacation and they knew that during the course of their research they

Fig. 1. The HeyStaks system architecture and outline recommendation model.
would use Web search as their pr imary source of information about what to
do and where to visit. Steve created a (private) search stak called “European
Vacation 2008” and shared this with Bill and friends, encouraging them to use
this stak for their vacation-r elated sear ches.
The HeyStaks Toolbar
The Stak-List
Tag, Share, Vote Actions
Create, Share, Remove Staks
Fig. 2. Selecting a new active stak.
Fig. 2 s hows Steve selecting this stak as he embarks on a new search for
“Dublin hotels”, and Fig. 3 shows the results of this search. The usual Goo gle
results are s hown, but in addition HeyStaks has made two promotions. These
were promoted because other members of the “European Vacation 2008 ” stak
had recently found these results to be releva nt; perhaps they selected them for
similar queries , or voted for them, or tagged them with related terms. Thes e
recommendations may have been pro moted from much deeper within the Google
result-list, or they may not even be present in Google’s default r e sults. Other

relevant results may also be highlighted by HeyStaks, but left in their default
Google position. In this way Steve and Bill benefit from promotions that a re
based on their previous similar searches. In addition, HeyStaks can recommend
results from other related public staks as appropriate, helping searchers to benefit
from the search knowledge that other groups and communities have created.
HeyStaks Promotions
Pop-up tag, share, vote icons
Fig. 3. Google search results with HeyStaks promotions.
Separately from the toolbar, HeyStaks users also benefit from the HeyStaks
search portal, which provides a social networking service built around people’s
search histories. For example, Fig. 4 shows the portal page for the “European
Vacation 2008” stak, which is available to all stak member s. It presents an ac-
tivity feed of recent search histo ry and a query cloud that makes it easy for the
user to find out about what others have been sea rching for. The search portal
also provides users with a wide range of features such as stak maintenance (e.g.,
editing, moving, copying results in staks and between staks), various search and
filtering tools, a nd a variety of features to manage their own search profiles and
find new search partners.
2.2 The HeyStaks Recomendation Engine
In HeyStaks each search stak (S) serves as a profile of the search ac tivities of
the stak members a nd HeyStaks combines a number of implicit and explicit

Frequently asked questions

Q1. What are the contributions in this paper?

In this paper the authors describe a novel and practical approach to Web search that combines ideas from personalization and social networking to provide a more collaborative search experience. The authors described how this has been delivered by complementing, rather than competing with, mainstream search engines, which offers considerable business potential in a Google-dominated search marketplace. 

Q2. What is the main feature of the HeyStaks search portal?

The search portal also provides users with a wide range of features such as stak maintenance (e.g., editing, moving, copying results in staks and between staks), various search and filtering tools, and a variety of features to manage their own search profiles and find new search partners. 

Q3. How many users have benefitted from the collaboration features of HeyStaks?

In turn, collaboration has begun to pay dividends for early HeyStaks users: 85% of users have benefitted from the search experiences of others and, on average, 34% of the time users are seen to select promotions that have originated from their peers. 

Q4. What was the effect of the idea of using connectivity information to rank search results?

In the late 1990’s the world of Web search was transformed by the idea of using connectivity information to rank search results, and within a few short years Google’s PageRank had rendered purely term-based approaches obsolete. 

Q5. What is the definition of a net producer?

To be clear a net producer is defined as a user who has helped more other users than they themselves have been helped by, where as a net consumer is defined as a user who has been helped by more users than they themselves have helped. 

Q6. How many users have adapted to the collaboration features of HeyStaks?

The authors have presented the results of a recent deployment that highlight how many early users have adapted well to the collaboration features offered by HeyStaks: most users create multiple search staks to store their search experiences and 70% of users share staks with others. 

Q7. What is the main reason for poor search productivity?

Poor search productivity is largely a result of vague or ambiguous queries [6, 8, 20], and there is considerable research on different ways to improve result selection and ranking. 

Q8. How many users were prolific stak creators?

A few users were prolific stak creators and joiners: one user created 13 staks andjoined another 11, to create a search network of 47 other searchers (users who co-shared the same staks). 

Q9. What is the main benefit of the HeyStaks search portal?

Separately from the toolbar, HeyStaks users also benefit from the HeyStaks search portal, which provides a social networking service built around people’s search histories. 

Q10. how many users are involved in the creation of useful search knowledge?

Unlike other forms of social media, where a minority of users (< 10%) participate in production, the authors have found that more than half of HeyStaks users are involved in the creation of useful search knowledge. 

Q11. What is the main idea behind the paper?

The authors believe that social (or collaborative) search techniques have the potential to have a similarly transformative impact on current Web search, and in this paper the authors have described the result of one research project in this area which has now matured in to a commercial venture.