Handbook of Research onWeb Log AnalysisBernard J. JansenPennsylvania State University, USAAmanda SpinkQueensland Universit...
Director of Editorial Content:    Kristin KlingerDirector of Production:           Jennifer NeidigManaging Editor:        ...
Chapter XXII      Analysis and Evaluation of the           Connector Website                                              ...
Analysis and Evaluation of the Connector Websitebring many people together who otherwise would         bone for connector ...
Analysis and Evaluation of the Connector Websitecommunities (Boase et al., 2006; Horrigan, 2001;        tHE cONNEctOr WEbs...
Analysis and Evaluation of the Connector Website    The Journal of Computer-Mediated Com-               communications and...
Analysis and Evaluation of the Connector WebsiteTable 1. Use and time commitment for connector websites              Short...
Analysis and Evaluation of the Connector WebsiteFigure 1. Connector website dimensions                                    ...
Analysis and Evaluation of the Connector Websiteresearch on connector Websites                          2007; Kim  Yun, 20...
Analysis and Evaluation of the Connector Websitetrade practically anything.” So far it has worked      only did eBay insta...
Analysis and Evaluation of the Connector Websiteby way of the Web. Now more than 14 million              of separation” ex...
Analysis and Evaluation of the Connector Websitetwenty-somethings, who by and large looked for         model, but focused ...
Analysis and Evaluation of the Connector Websiteto systematically evaluate Website traffic to de-               highly-pub...
Analysis and Evaluation of the Connector Websitenine months, May 2005 through September 2007.           viewed as a type o...
Analysis and Evaluation of the Connector Websitein September 2007. MySpace clearly separates                    nine month...
Analysis and Evaluation of the Connector Websitevariable (month), and y is the value of the depen-            be a “strong...
Analysis and Evaluation of the Connector Websiteconnectors fall within the “moderate” category.          level of core onl...
Analysis and Evaluation of the Connector WebsiteTable 6. How volatile are the connector websites in the sample? MAY 2005 t...
Analysis and Evaluation of the Connector Websitepresented in the preceding analysis give sugges-       sions and mountain ...
Analysis and Evaluation of the Connector Websitethe allure of exponential network growth rates is     a customer service m...
Analysis and Evaluation of the Connector Websiteit is possible that Flickr will establish itself as    communications or t...
Analysis and Evaluation of the Connector Websitecould organize around occupations, education,                             ...
Analysis and Evaluation of the Connector Website1.    Social Values and Tradeoffs: Connector                   als; family...
Analysis and Evaluation of the Connector WebsiteAcKNOWLEDGMENt                                       Boase, J., Horrigan, ...
Analysis and Evaluation of the Connector Website
Analysis and Evaluation of the Connector Website
Analysis and Evaluation of the Connector Website
Analysis and Evaluation of the Connector Website
Analysis and Evaluation of the Connector Website
Analysis and Evaluation of the Connector Website
Analysis and Evaluation of the Connector Website
Analysis and Evaluation of the Connector Website
Analysis and Evaluation of the Connector Website
Analysis and Evaluation of the Connector Website
Analysis and Evaluation of the Connector Website
Upcoming SlideShare
Loading in...5

Analysis and Evaluation of the Connector Website


Published on

Published in Handbook of Research on Web Log Analysis, (Hershey, PA: IGI-Global, 2009).

Published in: Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total Views
On Slideshare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Analysis and Evaluation of the Connector Website

  1. 1. Handbook of Research onWeb Log AnalysisBernard J. JansenPennsylvania State University, USAAmanda SpinkQueensland University of Technology, AustraliaIsak TaksaBaruch College, City University of New York, USA InformatIon scIence reference Hershey • New York
  2. 2. Director of Editorial Content: Kristin KlingerDirector of Production: Jennifer NeidigManaging Editor: Jamie SnavelyAssistant Managing Editor: Carole CoulsonTypesetter: Sean WoznickiCover Design: Lisa TosheffPrinted at: Yurchak Printing Inc.Published in the United States of America by Information Science Reference (an imprint of IGI Global) 701 E. Chocolate Avenue, Suite 200 Hershey PA 17033 Tel: 717-533-8845 Fax: 717-533-8661 E-mail: cust@igi-global.com Web site: http://www.igi-global.comand in the United Kingdom by Information Science Reference (an imprint of IGI Global) 3 Henrietta Street Covent Garden London WC2E 8LU Tel: 44 20 7240 0856 Fax: 44 20 7379 0609 Web site: http://www.eurospanbookstore.comCopyright © 2009 by IGI Global. All rights reserved. No part of this publication may be reproduced, stored or distributed in any form or byany means, electronic or mechanical, including photocopying, without written permission from the publisher.Product or company names used in this set are for identification purposes only. Inclusion of the names of the products or companies doesnot indicate a claim of ownership by IGI Global of the trademark or registered trademark.Library of Congress Cataloging-in-Publication DataHandbook of web log analysis / Bernard J. Jansen, Amanda Spink and Isak Taksa, editors. p. cm. Includes bibliographical references and index. Summary: “This book reflects on the multifaceted themes of Web use and presents various approaches to log analysis”--Provided bypublisher. ISBN 978-1-60566-974-8 (hardcover) -- ISBN 978-1-60566-975-5 (ebook) 1. World Wide Web--Handbooks, manuals, etc. 2. Web usage mining--Handbooks, manuals, etc. I. Jansen, Bernard J. II. Spink, Amanda.III. Taksai, Isak, 1948- TK5105.888.H3636 2008 006.3’12--dc22 2008016296British Cataloguing in Publication DataA Cataloguing in Publication record for this book is available from the British Library.All work contributed to this book set is original material. The views expressed in this book are those of the authors, but not necessarily ofthe publisher. If a library purchased a print copy of this publication, please go to http://www.igi-global.com/agreement for information on activating the librarys complimentary electronic access to this publication.
  3. 3. Chapter XXII Analysis and Evaluation of the Connector Website Paul DiPerna The Blau Exchange Project, USAAbstrActThis chapter proposes a new theoretical construct for evaluating Websites that facilitate online socialnetworks. The suggested model considers previous academic work related to social networks and onlinecommunities. This study’s main purpose is to define a new kind of social institution, called a “connec-tor Website”, and provide a means for objectively analyzing Web-based organizations that empowerusers to form online social networks. Several statistical approaches are used to gauge Website-levelgrowth, trend lines, and volatility. This project sets out to determine whether or not particular connec-tor Websites can be mechanisms for social change, and to quantify the nature of the observed socialchange. The author hopes this chapter introduces new applications for Web log analysis by evaluatingconnector Websites and their organizations.PrEMIsE a resulting social change that occurs quickly and exponentially (Gladwell, 2000).In February 2000, Malcolm Gladwell published How does tipping theory work? A diseasehis best-selling book The Tipping Point to much epidemic is the metaphor. According to Gladwell,controversy as well as popular acclaim. Gladwell’s social tipping happens as a viral process by whichthesis, building on the work of epidemiologists ideas, products, and messages spread and infectand social scientists before him, stated that a a population. Gladwell’s “Connector” is a keysocial “tipping point” is characterized by: (1) human agent in social epidemics. Connectors arethe contagiousness and subsequent “stickiness” important because they know a lot of people, andof an idea, product, or message; (2) small causes the people they know come from diverse socialleading to big effects and social change; and (3) networks and subcultures. Simply put, ConnectorsCopyright © 2009, IGI Global, distributing in print or electronic forms without written permission of IGI Global is prohibited.
  4. 4. Analysis and Evaluation of the Connector Websitebring many people together who otherwise would bone for connector Websites. To some degree,have little chance to make an acquaintance. If each Website allows for “social search” andthese people casually interact by exchanging ideas, “social networking”. It is an empirical questionproducts, or messages (relying on the first factor beyond the scope of this report to parse out toof contagiousness and stickiness), interactions what extent a Website is used specifically for oneshould exponentially increase across a defined purpose or the other. In general, connectors allowpopulation and lead to a social tipping point. users to create self-identifying profiles, while also A classic social networking model put forth empowering them to search for others based onmore than thirty years ago by Mark Granovetter needs, interests, mutual “friends”, contacts, orexplains one of the general processes driving other points of focus.tipping point theory. In an article published in In the mid-to-late 1990s, the first connectorthe American Journal of Sociology, Granovetter Websites were those emphasizing social search,explored how limited small-scale social interac- and more specifically, online dating (e.g. Match.tions (he termed “weak ties”) can lead to large- com), online trading and classifieds (e.g. Craig-scale phenomena such as diffusion of influence slist), and online auctions (e.g. eBay). A secondand information, social mobility, community generation of connectors gained national mediaand political organization, and social cohesion attention around 2002, offering more explicit(Granovetter, 1973). Weak ties tend to be low social networking options for professional/careermaintenance in terms of time, commitment, and networking (e.g. LinkedIn, Ryze), and for makingenergy. These relationships often provide a broad new friends through mutual friends or interestsrange of sources that are quick references for new (e.g. Friendster, MySpace, Facebook). In 2004information and opportunities. Gladwell’s Con- industry-leading companies like Google, Yahoo!,nectors enable the weak ties between people. MSN, and AOL started testing their own con- nector Websites to enhance their existing online communities.INtrODUctION The research on connector Websites is largely qualitative and restricted to social networks andToday many millions of Americans are utilizing online communities within Websites. Most re-“connector Websites” to serve as a proxy for searchers have focused their energies modelingGladwell’s Connector. The connector Website is the design of online communities, observinga proposed theoretical construct and is defined in online identity formation and social interactionsthis report. This type of Website is a new kind within one Website or a small number of Websites,of social institution, and its public availability and using social network analysis to measure thecoincided with the emergence of the Internet kinds of relationships in a community’s socialin the mid-1990s. A connector Website has the networks (Adamic et al., 2003; Barnes, 2006;capacity and function to provide contacts and Boyd, 2004; Boyd Heer, 2006; Donath Boyd,facilitate social exchanges between people, and 2004; Dwyer, 2007; Garton et al., 1997; Kollockeffectively build communities of users. It boosts Smith, 1999; Preece Maloney-Krichmar,timely and relevant interactions between individu- 2003; Rheingold, 1993; Rosen et al., 2003; Well-als while enlarging the scale of social exchange man Gulia, 1999; Wellman Hampton, 1999;processes, by way of online social search and Stutzman, 2006a; Stutzman, 2006c). In recentsocial networking. years, the Pew Internet American Life Project Social exchange applications (and technolo- has conducted large scale surveys offering valu-gies) collectively fortify the infrastructural back- able information about human behavior in online
  5. 5. Analysis and Evaluation of the Connector Websitecommunities (Boase et al., 2006; Horrigan, 2001; tHE cONNEctOr WEbsItELenhart Madden, 2007). However there is little MODELsocial science research on Websites, function-ing as social organizations and producing social In early October 2005, at the annual Online Com-networks and online communities. munity Summit, enthusiasm abounded for the When arguing the importance for studying potential of online collaboration, social networks,institutions in economics, legendary economist and community-building. Many in attendanceRonald Coase stated: “… it is the institutions viewed the Internet’s quickly evolving socialthat govern the performance of an economy, and applications as a boon for Website development.it is this that gives the [study of institutions] its Many attendees left the conference with an op-importance for economists” (Coase, 1998). Like- timism believing that online social networks andwise, this study assumes the general importance communities can do social good, whether throughof studying social institutions to the performance private, nonprofit, or public sector organizations.of a social system. In this chapter connector A new kind of Website had emerged allowingWebsites are treated as institutions whose social people to communicate, organize, and coordinatecontext – development, operations, culture, and with each other in new ways.governance – have real world implications for the The connector Website model is based onperformance of their respective social networks James Coleman’s reasoning that a person makingand online communities, as well as the larger even a limited social connection will affect thesocial system that is the Internet. status quo (Coleman, 1988). There are several The current social science research does not essential aspects when describing a connectortake the broad view, examining tangible implica- Website.tions of connector Websites. This chapter’s core The following is an operational definition:objectives are definition, discovery, exploration,and description. The chapter examines four • The connector Website provides a relativelyquestions: simple means of interaction for users who seek to offer or obtain goods, services, or1. What is the connector Website model? information.2. Which Websites have pioneered the con- • It is an intermediary offering peer-to-peer nector Website model? Web applications that collectively make3. Do connector Website trends demonstrate up an infrastructure for social exchange, exponential social change? networking, and diffusion processes.4. Do lessons of existing connector Websites • Over time, user-to-user interactions gradu- have future implications? ally generate a majority portion of the Website content, the regulation of which is By building on useful social theories, case stud- governed jointly between the host organiza-ies, survey research, and the snapshot reporting tion and the online community of users.of journalists, bloggers, and market analysts, the • Depending on the surrounding social andauthor hopes to better understand Websites whose economic conditions, as well as site designbroad implications are unknown. This study and development, the connector Websiteshould contribute to Web log analysis methodol- is capable of facilitating the discovery andogy, assessing the impact of a connector Website coordination of context-based communica-in terms of its aggregate effects. tions and transactions.
  6. 6. Analysis and Evaluation of the Connector Website The Journal of Computer-Mediated Com- communications and transactions are relativelymunication (JCMC) recently published a special simple. Specifically, connectors allow users tothematic issue on social network Websites, and look for others by targeting online profiles eitherso it is important to make a distinction between through mutual contacts or customized searchthe terms “connector Website” and “social net- parameters. A user can also perform a searchwork site”. JCMC guest editors Danah Boyd and on one or more preference criteria, or keywords,Nicole Ellison define a “social network site” in specifying which qualities he or she most desirestheir insightful introduction: in another user. Depending on the level of detail, a person can search for others based on one cat-We define social network sites as Web-based egory (e.g. zip code) or multiple categories (e.g.services that allow individuals to (1) construct gender, career interest, favorite sports, favoritea public or semi-public profile within a bounded movies, likes to cook, etc.). Interactive classifiedssystem, (2) articulate a list of other users with and discussion boards are other applications thatwhom they share a connection, and (3) view and facilitate interactions. These early Website ap-traverse their list of connections and those made plications were rooted in social search. Severalby others within the system. The nature and no- years later, a more explicit social networkingmenclature of these connections may vary from application added another dimension to the con-site to site. (Boyd Ellison, 2007) nector Website. Second, a connector Website serves as a key Social network sites are viewed here as a subset intermediary for its users. It serves this functionof the connector Website universe – they are the by actively bridging users who act as “consum-focus for data analysis later in this chapter. Gener- ers” with those users who are “producers”. Inally speaking, the definition for a social network an idealized model, consumer-users create thesite is structural. Implementation of a particular demand for information and other goods, andkind of social networking platform is necessary the producer-users supply the timely and rel-for a Website to be categorized as a social network evant information or goods. For example, eBaysite. It should “enable users to articulate and make or Craigslist can easily connect a user wantingvisible their social networks” (Boyd Ellison, a U2 poster or a GE microwave with other users2007). The six degrees computing concept, which who can supply these goods. Facebook, MySpace,Friendster made popular by 2004, is the origin for and Classmates.com are intermediaries for mak-publicly visible social networks on Websites. ing human contacts. Connectors tend to tap into For better or worse, the connector Website’s a previously underserved supply and demanddefinition is broader, containing both structural market, mediating frequent and new exchangesand operational elements. Connectors will have within its online community.searchable profiles, including but not confined Third, most connector Website content isto “six degrees” social search. Other conditions generated by the online community. A usefulrequire some combination of structural features metaphor is a professional conference. The con-such as social networking platforms, blogs, nector is the host of an online “conference” thatforums, RSS feeds, tags, wikis, and widgets never adjourns, running twenty-four hours a day,with observable social processes such as visits, seven days a week. The connector Website enablessearch, networking, exchange, diffusion, and the introductions and establishes the structure,governance. goals, usability, and social norms and values for A little more elaboration about connector the conference. However the attendees (users)Websites might be useful. First, interpersonal significantly contribute and add value to the
  7. 7. Analysis and Evaluation of the Connector WebsiteTable 1. Use and time commitment for connector websites Short Term Medium Term Long Term - advice - collaboration - collaboration - affirmation - focus groups - organization - commerce - interviews - peer support - corroboration - mobilization - research - elaboration - peer support - self-expression - peer support - research - working relationships - reference - self-expression - research - working relationships - resources - self-expression - tradingproceedings (Website content) as time moves on. once social norms are clear and well-established.This could mean establishing new working rela- This means the users collectively take on moretionships or cultivating existing ones, presenting responsibility for enforcing Website and commu-products and personal/professional information, nity values with respect to the user-posted con-advancing reputation, and offering any number tent. Accountability systems fueled by feedbackof other informal services. and reputation-building technologies (providing The connector Website’s content responsibili- cues like ratings and reviews) signal importantties alternate between its host organization and its information about a user and his or her content.group of users. Initially, right after Website launch, By implementing some kind of accountabilityit is important for the Website to supply content system a connector subtly structures online in-and the “rules of the game” for the community. centives in such a way as to guide user behavior.Online tutorials help users to learn to navigate and There is a cautionary note, however. Evidenceuse the Website. The connector ideally supplies exists of connectors applying heavy-handedonline material to spark discussions or ideas for top-down regulation, and as result, Websitesother interactions. Managing user expectations is have lost users. To some degree, connectors withalso critical. As connectors mature over time, a some self-regulation have user leaders or mavenssignificant amount of content begins to be directly who volunteer to enforce the connector’s normsgenerated by the online community – examples and values. These members function like onlineare user profile pages, blogs, tags, wikis, forums, neighborhood watchdogs.“real simple syndication” (RSS) feeds, ratings, The connector Website adds substantial valuereviews, classifieds, and replicable widgets. There to online communities when executing two pro-are also connector-community interactions such cesses: discovery and coordination. Connectorsas group blogs, group wikis, chat sessions, sur- make discovery more powerful and accurateveys, and polls. The connector may offer timely through social search and trust-building applica-and relevant classifieds, Website links, blogs, and tions. Joe Cothrel, an online community expert atinterviews. User interactions drive the Website’s Lithium Technologies, has suggested that “withdynamics, and in the process, further build the these tools, a user not only finds that someoneonline community. has something they need – [he or she] also finds Finally, online communities gradually become out something about the quality of that productjointly self-regulating with the host organization and the likelihood this party will deliver it” (J.0
  8. 8. Analysis and Evaluation of the Connector WebsiteFigure 1. Connector website dimensions CONNECTOR WEBSITE COMMUNICATIONS TRANSACTIONS SOCIAL SOCIAL SEARCH NETWORKING DISCUSISION BOARDS INSTANT MESSENGER RATINGS REVIEWS MEMBER PROFILES CLASSIFIEDS RSS FEEDS FORUMS WEB MAIL CHAT BLOGS WIKIS TAGSCothrel, personal communication, 2005). In terms on common needs, interests, and priorities. Theof coordination, connectors are more efficient model seizes on the low transaction costs of usingthan other online community Websites because the Internet. It also has the potential to forge weakthey have the capacity to network specialized (but important) contacts and to develop lastingcommunications and transactions for users in relationships. The connector empowers the indi-specific situations (DiPerna, 2006; Stutzman, vidual. The frequency, diversity, and informality2006b). Often social search and social network- of online social exchanges expose people to newing applications are mixed together on connector perspectives and experiences. A diffusion processWebsites and previously have been regarded by could lead to larger scale, possibly exponentialand large as a single application. social change. This study attempts to be a start- Why is a connector Website appealing? The ing point for new research on Websites as socialconnector saves time and energy for people, of- organizations and online social networks andfering effective ways to link up with others based communities as social systems.
  9. 9. Analysis and Evaluation of the Connector Websiteresearch on connector Websites 2007; Kim Yun, 2007; Liu, 2007; Lange, 2007). To date most studies on connectors have focusedThe existing research literature on Websites on one or two Websites at a time.– meeting this chapter’s definition of a connector However, Eszter Hargittai (2007) offered aWebsite – is sparse. Generally, projects have been compelling comparative analysis when lookinggrounded in theoretical and formal modeling, for systematic differences between people whosocial network analysis, or case study. Univer- use social network sites and those who do not usesity of California-Berkeley, Michigan State, and them. While analyzing survey data, Hargittai wasMassachusetts Institute of Technology research- able to break out distinct user characteristics forers have described Website applications within people on Facebook, MySpace, Friendster, andconnectors, namely Friendster (Boyd, 2004; Xanga. This chapter later provides a comparativeBoyd, 2006; Boyd Heer, 2006; Donath Boyd, traffic analysis for connector Websites.2004; Boyd Ellison, 2007). They have givenan early overview of social Websites; conductedethnographic fieldwork to describe the simultane- PIONEErING cONNEctOrous evolutionary patterns of a connector’s online WEbsItEscommunity and its Web-based social applications;and studied the roles of online profiles and identity Connector Websites target both general andfor communications and social networking. particular communities whose members are The interest in examining connector Websites, interested in individual expression, trading andparticularly social network sites, has exploded in auctioning of goods, matchmaking and dating,the last couple of years. Fred Stutzman (2006a; building social/friend networks, professional/ca-2006c) has uncovered interesting trends with reer networking, civic organization, event plan-respect to Facebook registration and self-identi- ning, and other social activities. For nearly allfication behavior. Nicole Ellison, Charles Stein- of these Websites, word-of-mouth referrals havefield, and Cliff Lampe (2007) have explored the been an effective marketing tool. Connectors’formation and maintenance of social capital on registration numbers and business activities areFacebook. Larry Rosen (2006), a psychologist sizable – they have social exchange applicationsbased at California State University-Dominguez that appeal to tens of millions of people, as wellHills, recently studied the behavior of Los Angeles as investors.area MySpace users. In August 2006, Cornell Uni-versity researchers presented a conference paper First Generation connectorsdiscussing the challenges of collecting and analyz-ing longitudinal data on online social groups and About ten years ago, connectors established ancommunities, specifically investigating the con- Internet presence on several fronts in social searchnector Website LiveJournal and a smaller online – online auctions, classifieds, and online dating.conference community (Backstrom et al., 2006). In the first case, transactions-based connectorsThey used social networking analysis and formal surfaced on the Internet in 1995. eBay (www.ebay.modeling techniques to consider the ways in which com), built by Pierre Omidyar and Jeff Skoll, iscommunities in online social networks grow over the alpha Website of this group. A decade old andtime. In the previously mentioned special issue of now boasting more than a hundred million activeJCMC, Internet researchers investigated the cul- members, the service is known to most Ameri-tures and online behaviors exhibited on MySpace, cans. eBay’s mission is “…to provide a globalBlackPlanet.com, Cyworld, and YouTube (Byrne, trading platform where practically anyone can
  10. 10. Analysis and Evaluation of the Connector Websitetrade practically anything.” So far it has worked only did eBay install a system to valuate transac-to unparalleled success on the Internet. Millions tion items, but it also systemized the valuation ofof items are listed in an online auction format. its community members.A seller can upload photos and other descriptive Craigslist (www.craigslist.org) is a conceptualmaterials for whatever he or she would like to sell. cousin of eBay, though without the nested ac-That person sets an initial bidding price and period countability system. Craigslist uses an interactiveof time for the auction. Many sellers also can set online classifieds format rather than eBay’s onlinea fixed price to forego the auction process. On the auctions format – a simple online bulletin boardother side of the transaction process, buyers either display, which is highly functional, and is easy tocasually browse or selectively search through use. The searchable classifieds are within a chosenitems for their desired consumption. Today eBay city of interest (in contrast to employing a profileserves as an auction house and shopping mall. If or keyword matching and sorting method likeever the cliché might be appropriate, Omidyar other social search applications). Craigslist offersand Skoll were able to deliver an idea whose time an organized directory of topics, and a familiarhad come – a hyper efficient trading community classifieds display to enable a swift user learn-guided by online social search applications and ing curve. Users post announcements or submitaccountability mechanisms. responses with ease, and so the social exchanges One of the many innovations at eBay is the are quick and direct. Unlike eBay, Craigslist doesway it provides information to sellers and buy- not use feedback or ratings mechanisms to builders about each other. The Website relies heavily trust. Most transactions tend to be conducted face-on a feedback system that builds user (buyer or to-face whether someone wants concert tickets, anseller) reputation, and as a result, installs a use- apartment, jobs, or other items and services.ful accountability system. After a transaction, At about the same time eBay and Craigslistthe buyer and the seller are allowed to rate the started out their paths on the Internet, online dat-person or some other entity on the other end of ing Websites started to gain social acceptance.the transaction on a one-to-five star scale. Accu- Match.com (www.match.com) has been a premiermulating these ratings, good or bad, contribute online dating Website since 1995. Today theto additional signals about a user’s trustworthi- Website boasts that millions use its social searchness. This important information establishes a applications for dating and relationships. Match.known value, which is important when making com has been successful because it attracts highlyany transaction. People are more likely to make motivated and engaged people – those who couldrational choices than irrational ones, and so they be lonely, looking to find dates, or hoping to startwant as much information as possible about whom a romantic relationship.and what they are dealing with in order to base an Classmates.com (www.classmates.com), alsoeconomic decision to serve their interests. launched in 1995, served implicit social network- Maybe one of eBay’s largest social contribu- ing goals, but it technically relied on a socialtions is how it has helped the general Internet search application. This early connector’s statedpopulation to recognize new norms in social ex- mission was (and still is) “…connecting millionschange. eBay is not the only Website responsible of members throughout the U.S. and Canada withfor this phenomenon, but it likely can share the friends and acquaintances from school, work, andcredit. The Website service constructed an online the military.” Classmates.com could be consideredsystem of “trust cues” about its members, and in a precursor to a later generation of connectorthe process, they possibly have helped establish Websites because it helped familiarize U.S. In-ratings and feedback banks as Web currency. Not ternet users with the notion of social networking
  11. 11. Analysis and Evaluation of the Connector Websiteby way of the Web. Now more than 14 million of separation” experiment (Dodds, Muhamad,different people visit Classmates.com each month Watts, 2003; Milgram, 1967). The process is(MarketWatch, 2006). Social search made inroads simple and repetitive to find an “nth degree” friend.for online trading, dating, and reunion, but a few When a user starts at his or her own profile homemore years would go by before discrete social page, this person can readily see his or her firstnetworking tools emerged on Websites, offering degree friends. It is then possible to go to a firsta new way for social exchange. degree friend’s profile and see all of that person’s first degree friends, the originating user’s secondsecond Generation connectors: degree friends. The process can begin again withEmergence of social Network sites a second degree friend’s profile in order to find third degree friends, and so on.In 2003-2004, Friendster (www.friendster.com) Friendster makes social search very easy,ushered in a new wave of connector Websites connecting people by particular demographicsand garnered a lot of publicity. Friendster offered and keywords. For example, if a person wants toa novel social networking application that was find others who have liked reading the Da Vincipartly based on dating Websites’ use of searchable Code or someone who happens to follow theonline profiles, as well as the six degrees email Pittsburgh Steelers, Friendster is a sure bet fornetworking technology introduced in 1997 by matching with other like-minded people basedSixdegrees.com. Friendster allowed users to make on those parameters. The Website enables a hightheir online friends publicly visible, as well as view likelihood of finding other people with a mutualand even contact friends of friends in the order of interest in a short amount of time. Because offirst, second, and third degrees. Friendster’s rapid this creative mixture of social search and socialuser growth inspired many copycat Websites. Big networking applications, after three years ofInternet companies, like Yahoo!, Google, and service, Friendster publicized on its Website thatMicrosoft – entered the online social network it had more than 20 million registered users. Insite space by 2004. an effort to broaden its appeal, Friendster began There were fascinating upsides and downsides to offer blogging, classifieds, bookmark sharing,in Friendster’s early story. To its credit, Friendster and other interactive capabilities.had been used for purposes ranging from online Competition and internal strife switcheddating to meeting people based on personal fortunes for Friendster. After an initial splash ofbackground, location, interests, or pre-existing success, the connector Website began having afriendships. On the other hand, Friendster did number of nagging problems. The Website earnednot have much in the way of original content, nor the unwanted reputation for technical miscalcu-did it seem to have a core understanding of one lations and poor customer service. Friendster’sor more distinguishable groups of its users. As a online infrastructure was not ready for its earlyresult, when faced with competition, the connector surging user growth rates. People complainedhad to scramble to brand itself. about major lag times between making profile Friendster includes many of the features of edits and those edits being officially posted ononline dating connectors, but its main appeal has the Website. Another problem was Friendster’sbeen functionality. The Website offers searches lack of brand or theme. It became a lot like afor people or groups across an unlimited number mall food court, a purely functional place thatof social networks or topic categories. On Friend- tries to offer many different things to manyster, contacting the friends of friends is based in different people. Demographics emerged as thepart on Stanley Milgram’s famous “six degrees most defining character of the Website – urban
  12. 12. Analysis and Evaluation of the Connector Websitetwenty-somethings, who by and large looked for model, but focused on a more targeted audiencedates or competed with each other to create the – college students. Mark Zuckerberg, Dustinlargest friend network. At best, this was passive Moskovitz, and Chris Hughes, all students atbranding – Friendster did not actively target these Harvard University, launched the Website in earlygroups of users. 2004. Danah Boyd has described Facebook as a Over time, Friendster’s functionality did not “closed” network – memberships restricted onlyseem enough to earn user loyalty. Many users de- to those people with academic email addressesfected to newer connector Websites like MySpace and searching only locally within schools and notand Facebook. Competitors could provide the across schools (D. Boyd, personal communication,same social exchange applications, and to grow- 2005). In just over two years, Facebook was ableing online communities, while also doing more to claim at least 80% participation on nearly everyto engage on themes or targeted demographics. major college campus in America.Friendster accomplished two necessary ingredi- LinkedIn (www.linkedin.com) also featuresents for success – early entry into a new industry social networking capabilities and uses an evenand offering a disruptive technology and innova- more restricted network model for nurturing itstion – but the Website to this day continues to online community. LinkedIn facilitates businesssearch for competitive advantage and long term and professional networking. The company de-sustainability. scribes its connecting protocol as “…users can be Over the past few years, MySpace (www. approached [only] if one of their trusted contactsmyspace.com) and some other competitors capi- forwards a contact request to them because theytalized on Friendster’s problems. On the surface, believe it is an opportunity their contact will ap-MySpace started out similar to Friendster – it was preciate.” In essence LinkedIn constrains accessa connector Website primarily based on social to its members. This is quite more restrictive thansearch and social networking applications for the MySpace, Friendster, and other more social orpurpose of context-based communications. At one friends-based Websites. Reid Hoffman, foundertime the demographic of the two sites were alike, of LinkedIn, converted a common offline socialtargeting teens and twenty-somethings. However process into an online process. Both LinkedIn andMySpace went a step further than Friendster Facebook have taken steps to structure their on-by seeking out a niche community of users by line communities keeping in mind certain needs,specifically focusing on music and popular enter- interests, and priorities, while establishing largetainment interests. The Website enabled people to and somewhat closed social networks.discover music online in the same way they find To sum up this section, first and second genera-out about music in person – through friends. At tion connector Websites clearly set out to enablean early stage, MySpace offered file downloads, weak ties between users, first by social search,direct connections to bands, testimonials and and then by phasing in social networking. Fol-ratings, and career networking capabilities for lowing the successes of pioneering connectors,musicians. At first a seeming benign derivative many of the 2005 and 2006 startup connectorsof Friendster, MySpace catapulted way ahead of embedded social networking applications intoits predecessor in terms of membership growth, their platforms. Given the premise stated earlieruser log time on the Website, and user satisfaction. that a connector generally enables social processesOn July 11, 2006, it was reported that MySpace that can lead to exponential change, we shoulddethroned Yahoo! as the most heavily visited expect these Websites to produce evidence ofWebsite on the Web. exponential trends. Longitudinal Website traffic Like MySpace, Facebook (www.facebook. data and user growth rates should be adequatecom) also utilized Friendster’s social networking measures. Little research has been conducted
  13. 13. Analysis and Evaluation of the Connector Websiteto systematically evaluate Website traffic to de- highly-publicized “Web 2.0” phenomenon, a termtermine whether or not a Website demonstrates popularized by Tim O’Reilly and the Internet tradeexponential change over time. media for peer-to-peer-oriented Websites, which Can user traffic data for a connector Website emphasize interaction, collaboration, and user-offer evidence of emergent online social structures generated content. A panel of comScore Mediaand behaviors leading to online communities? Metrix Website traffic data (U.S. only) is used forThis question is explored for the remainder of longitudinal analysis. Observations are restrictedthis study. by the availability of data for comScore-selected connectors that have, to some degree, installed social networking and social search applications.DAtA MEtHODs Connectors also had to meet the following condi- tions: (1) the Website’s online community gener-Monthly Website traffic data for thirteen con- ates a substantial amount of content, mostly bynector Websites are analyzed for the following profiles; and (2) the Websites launched before Julyanalysis. Most of these connectors are part of the 2005. The observed time period spanned twentyTable 2. Selected connector Websites NAME Online since* Social Search/Networking for** CLASSMATES.COM Jan-1995 Maintaining School, Work, Military Connections LIVEJOURNAL Mar-1999 Blogging – Ratings XANGA Nov-1999 Blogging; Exploring/Sharing Photos FRIENDSTER Apr-2002 General / Pop Culture Interests LINKEDIN May-2003 Business Networking (in more than 130 industries) MYSPACE Sep-2003 General / Pop Culture Interests HI5 Dec-2003 General / Latin America, Europe, North America, Australia Converting Offline Networks Into Online Networks (Google ORKUT Jan-2004 affiliate) FACEBOOK Feb-2004 College and High School Students Folksonomy/Tagging – Organizing/Sharing/Suggesting FLICKR Mar-2004 Photos and Images TAGGED Oct-2004 Tagging – Exploring/Sharing Photos, Videos, Bookmarks WINDOWS LIVE SPACES Dec-2004 Blogging, Sharing Photos United Kingdom – Teens – Schools, Colleges, and Music BEBO Jan-2005 Interests* Information obtained by following this sequence of sourcing: direct email with Website’s media relations; on the websiteitself; doing a Google search using “[website name] launched” and then verifying across multiple sources.** Information inferred from the website’s main webpage, “About Us” webpage, and (if applicable) Wikipedia entry.
  14. 14. Analysis and Evaluation of the Connector Websitenine months, May 2005 through September 2007. viewed as a type of social institution. This analysisThe unit of analysis is the connector Website, and is strongly suggestive, but not conclusive, sincethe key dependent variable is “Monthly Unique there is a relatively small number of connectorVisitors” – estimated counts of individual users Websites contained in the dataset. In any case itwho view a Website in a particular month. should be possible to learn more about longitudinal This study is an exercise in Web log analysis. trends on thirteen popular connector WebsitesAdmittedly, monthly unique visitors could be one and see if traffic trends have implications forof the most basic pieces of data for doing such work. social change.However it is possible to glean insightful findingsfrom even the most basic type of measurement,especially when a given body of research for these cOMscOrE MEDIA MEtrIXnew Websites is so young. The data also allow ANALysIsfor interesting derivative measures for extendinganalysis and understanding. Table 3 lists the connector Websites in rank order Do connector Websites offer evidence of of monthly unique visitors (hereafter, traffic)exponential social change over time? This is thekey empirical question for this chapter. This studyconsiders month-to-month traffic changes as a Table 3. Monthly unique visitors on connectorsnapshot of social change for a given Website’s Websites.* SEPTEMBER 2007online community. The comScore Website traf- CONNECTOR WEBSITE No. Unique Visitors (000)fic data lead to traffic trend estimations for each MYSPACE 68,449of the thirteen connector Websites. At the end ofthis section, a judgment is made on whether or FACEBOOK 30,601not a given connector demonstrated exponential FLICKR 13,150change in the time period May 2005 throughSeptember 2007. The determination is based on CLASSMATES.COM 13,084the following observations: WINDOWS LIVE SPACES 9,7631. Percentage change over the time period, BEBO 4,389 May 2005 to September 20072. Traffic trend estimation with respect to an LIVEJOURNAL 4,185 exponential function; XANGA 3,1233. Website volatility4. Website age and maturity HI5 2,833 LINKEDIN 2,386 A couple of cautions and limitations should bementioned at this point. The following Web log TAGGED 1,741analysis is exploratory and descriptive - causality FRIENDSTER 1,645cannot be asserted using the methods employedhere. The study is unable to distinguish between ORKUT 503the characteristics attributed to the host organi- U.S Internet Usage 181,858zation from those of the online community ofusers. The Website organization and the online * Data provided by comScore Media Metrix: http://www.community are a single symbiotic entity and comscore.com
  15. 15. Analysis and Evaluation of the Connector Websitein September 2007. MySpace clearly separates nine months. Flickr and LinkedIn are standoutitself from the pack, having more than twice the examples. Flickr increased its traffic more thannumber of users when compared to Facebook, its thirteen times its baseline. LinkedIn shot upnearest competing connector in this sample. More more than eleven times. Interestingly, affinitythan 1 out of 3 American Internet users visited social network site rivals MySpace and FacebookMySpace in the last observed month. Facebook demonstrated identical growth rates (339.4%)claims roughly 17% of U.S. Internet users. Flickr over the observed time period. Four connectorsand Classmates.com attract about 7%. Moving lost users during the time period, three of themdown the table, the connectors from Windows Live (Classmates.com, Xanga, and LiveJournal) areSpaces to LinkedIn attract between 5% and 1% of the oldest in the sample. Both of these connec-the U.S. Internet usage market. Tagged, Friendster, tors also maintained an average monthly trafficand Orkut are beneath the 1% threshold. growth rate above 10%. In Table 4 some connector Websites illustrate At this point, the study models each of thelarge total growth and impressive average monthly connector Website trends with respect to an ex-growth rates. Eight Websites at least doubled ponential function, using the equation: y = ae bx.their monthly traffic over the course of twenty In the equation x is the value of the independentTable 4. Which are the fastest growing connector websites?* MAY 2005 to SEPTEMBER 2007 Total Period Average Monthly WEBSITE % Traffic Change % Traffic Change FLICKR 1,324.0 % 11.2 % LINKEDIN 1,116.9 % 12.9 % ORKUT 343.3 % 8.7 % FACEBOOK 339.4 % 6.0 % MYSPACE 339.4 % 5.6 % TAGGED 282.2 % 7.4 % WINDOWS LIVE SPACES 211.0 % 4.6 % BEBO 118.0 % 4.8 % FRIENDSTER 29.6 % 2.2 % HI5 -4.3 % 2.0 % CLASSMATES.COM -22.9 % -0.1 % LIVEJOURNAL -43.4 % -1.4 % XANGA -62.8 % -2.3 % U.S. Internet Usage 9.9 % 0.3 %* Raw data provided by comScore Media Metrix: http://www.comscore.com
  16. 16. Analysis and Evaluation of the Connector Websitevariable (month), and y is the value of the depen- be a “strong” fit; .5 ≤ R 2 .8 to be “moderate”dent variable (monthly unique visitors). The value fit; and 0 ≥ R 2 .5 to be a “weak” fit. Six ofe (approximately 2.7182) is the base of natural the trends correlate nicely with its exponentiallogarithms. A statistical program calculates the model, indicating strong fit. Two of them – Flickrvalues of a and b that best fit the connector’s data. and LinkedIn – displayed staggering growthIf the observed data points closely correlate to over twenty nine months. Orkut, Facebook, andthe newly generated exponential curve, then this MySpace all more than tripled, but also impor-pattern (illustrated by a high R 2 value) indicates tantly, showed modest downturns like the otherthat exponential growth does a good job of ex- two higher growth Websites. Xanga is the onlyplaining a connector Website’s trend during the Website indicating what seems to be very seri-time period. ous decline. Tagged, Windows Live Spaces, and Table 5 summarizes the trends and the expo- Bebo also demonstrated solid growth spurts innential function tests. Appendix B lays out the this period, but the fluctuating rhythms of theirtrend graphs for the connector Websites in the trends likely hurt the correlation to their respec-sample. To err on the side of caution, R 2 ≥ .8 to tive exponential models. As a result, these threeTable 5. Does the connector Website demonstrate exponential change? Based on trend estimation, did the connector website demonstrate exponential change over the observed time period? (Y/N) MAY-05 to SEPT-07 Goodness of Fit WEBSITE Final Estimation % Traffic Change Exponential Function (R 2) FLICKR 1,324.0 % 0.9208 = STRONG Y LINKEDIN 1,116.9 % 0.9219 = STRONG Y ORKUT 343.3 % 0.8818 = STRONG Y FACEBOOK 339.4 % 0.9468 = STRONG Y MYSPACE 339.4 % 0.8794 = STRONG Y TAGGED 282.2 % 0.6013 = MODERATE N WINDOWS LIVE SPACES 211.0 % 0.5596 = MODERATE N BEBO 118.0 % 0.7635 = MODERATE N FRIENDSTER 29.6 % 0.1384 = WEAK N HI5 -4.3 % 0.1792 = WEAK N CLASSMATES.COM -22.9 % 0.3845 = WEAK N LIVEJOURNAL -43.4 % 0.3493 = WEAK N XANGA -62.8 % 0.8289 = STRONG Y
  17. 17. Analysis and Evaluation of the Connector Websiteconnectors fall within the “moderate” category. level of core online community and user loyalty.The other Websites did not show a dramatic per- Admittedly, this is mostly speculation. There iscentage growth nor did they fit particularly well a solid pathway for future research beyond thewith their exponential models. Seven connector range of this study.Websites do not pass this empirical test. Website Age and MaturityTraffic Volatility The age of the connector Website appears associ-Trend volatility might explain in part why Tagged ated with traffic growth differences. There is amisses the cut, and why MySpace, Facebook, moderate negative association (r=-.38) betweenand Orkut (showing only slightly greater relative a Website’s age (in months) and its percentagegrowth) fit their exponential models. Table 6 lists growth over the time period – meaning that olderthe connector Websites in ascending order, from age may bring slower growth. This is not a bigleast to most volatile. Volatility is defined here surprise. It is quite possible the older Websitessimply as the variance of a Website’s “monthly boomed at an earlier time before the observedpercentage change” over the twenty nine month time period. A maturation period may be atime period. We see LinkedIn and Orkut are at the characteristic for these Websites as they approachbottom of the list, having the steepest upturns and long term sustainability. It is also interesting todownturns on their respective month-to-month note that there is a moderate negative associationrates of change. It is not clear how volatility af- (r=-.32) between a Website’s age and its volatilityfects connector Websites. Orkut and LinkedIn, across the time period – meaning that older sitesstill relatively small in total users and achieving may experience less severe shifts in monthlyamong the highest growth rates, exhibit the most rates of change. It is intuitive to think that atrend volatility. On the other hand MySpace and connector’s host organization would eventuallyFacebook, the two most populated connectors, get past some of the early bumps and bruises thathave relatively moderate growth rates and are come as a start-up entity, leading to less severelow volatility. It’s interesting that LiveJournal, the traffic swings.poorest performing connector, also sits at the topof the table with a low volatility score. It may bethat volatility is the wild ride host organizations DIscUssION IMPLIcAtIONsneed to experience before reaching some stabiliz-ing threshold of monthly traffic. One basic finding stands out. Connector Websites In terms of growth and numbers, volatility experience exponential changes in Website trafficmight be a necessary condition to become a over time. At least five out of the thirteen connec-successful connector Website. High to moderate tors showed rapid growth, and one other displayedvolatility could be a characteristic at start-up and substantial shrinkage. So six connectors havewhile a Website escalates operations and activities. produced rather unexpected social epidemics inThough volatility is probably viewed as a liability terms of huge gains (or loss) in user traffic.after some kind of scale has presumably been As a side note, it would seem impossible to sus-reached, either in terms of community-building tain the kind of Flickr or LinkedIn traffic growthor organizational capacity. Following this line in the long-term. Take Flickr as an example. Theof thought, low to moderate volatility is prob- site is approximately 43 months old. Hypotheti-ably a characteristic of the most highly visited cally, if the Website kept escalating at the currentWebsites where they should have earned some pace of about 500% per year, it would approach0
  18. 18. Analysis and Evaluation of the Connector WebsiteTable 6. How volatile are the connector websites in the sample? MAY 2005 to SEPTEMBER 2007 WEBSITE Volatility* MYSPACE .0039 LESS VOLATILE WINDOWS LIVE SPACES .0115 LIVEJOURNAL .0116 FACEBOOK .0124 CLASSMATES.COM .0177 FRIENDSTER .0247 XANGA .0267 FLICKR .0308 BEBO .0440 HI5 .0531 TAGGED .0641 ORKUT .0657 MORE VOLATILE LINKEDIN .0872 U.S. Internet Usage .0001*Volatility defined here as the variance of “monthly percentage change” for a website over the twenty nine month timeperiod.1.6 billion unique monthly users in September Why does it look like some connector Websites2010 – that would be nearly 1 out of 5 people on illustrate exponential change, and others do not?the planet! Eventually growth has to decelerate. The average age of the connectors that exhibitTherefore it seems to be important for any con- positive exponential growth is approximately 47nector Website’s host organization to forecast its months. Based on their large online constituenciesmarket (or online community scale) at the early and relative “maturity”, it is a reasonable guessstages of the connector’s operations. Over time, that Classsmates.com, LiveJournal, and Xangasustaining a level relative to the forecasted scale (all three launched before 2000) may have shownmight be more prudent than seeking constant rapid sizable growth long before the time period in thisgrowth. This is not to understate the difficulty of analysis. There is anecdotal evidence that Friend-making such projections and calculations. Quite ster hit hyper growth in 2003-2004. It would bethe contrary, for a general service like Flickr (i.e. interesting to test whether or not there is a clearphoto sharing), forecasting for scale will likely maturation process for these Websites and theirbe difficult. host organizations. The moderate correlations
  19. 19. Analysis and Evaluation of the Connector Websitepresented in the preceding analysis give sugges- sions and mountain bikes. Sustainable connectorstive evidence on the importance of time and age actively address privacy and security dangers, andwith respect to exponential trends. they are committed to quality control. A mostly self-regulating online communityWhat do the Lessons of Existing is the desired means for enforcing Website rulesconnector Websites Imply for Future and norms. However, online vigilantism is anstartups? unintended consequence for such bottom-up enforcement of the community’s norms andConnector Websites have to contend with a num- values. eBay has been struggling with this issueber of issues. First, personal security and fraud over the past few years. eBay users have takenare serious issues. A host organization’s need to accountability measures into their own handsmaintain reliability and trust with its online com- mainly by alerting buyers that a listing is fake.munity, and any direct or perceived breach can be eBay encourages reporting fake listings to thedeath to the community. Connectors are virtual company, but they do not approve of user-to-open markets for social exchanges of information user alerts or self-styled sting operations. Theseor other goods and services, and they largely rely actions may undermine eBay’s accountabilityon self-policing and reporting of Website abuses. system, which is the primary signal for a givenNot long ago Friendster was notorious for its users user’s reputation and trustworthiness. The les-creating fictional profiles and even defamatory son here is that extreme regulation, bottom-upprofiles, also known as “fakesters”. Fortunately or top-down, is unhealthy for a connector’s longnew Web services are becoming available to term development. Online security violations andenable connectors to conduct verifications for fraud take many forms, but there is no doubt thatonline profiles as a means to prevent fakesters, quality control of user-generated content will beharassers, and underage users. a major concern for connectors. As time passes, A lot of connectors ask for photos to go with it is conceivable that host organizations adapt byonline profiles, and sometimes these photos encouraging online community rules, norms, andare quite explicit. Match.com has a fairly strict self-regulation, contributing to the maturationscreening process that mandates a waiting period period suggested in the study’s analysis.before photos are posted. Employees scrutinize Second, a key challenge for connector Websitessubmitted photos as a means of quality control. So is to attract a great quantity of users that provide anthe problem is resolvable, albeit likely requiring expansive range of perspectives and experiences.more labor and cost. For some connectors, photos Otherwise the utility of the connector diminishes.may not be necessary for facilitating interactions For example, if someone is an elementary schoolamong users. The creation of an avatar is a good principal looking for information in order to adoptexample of side-stepping the pitfalls of providing a new reading program, it would be most valu-profile photos. Users create an animated image to able if there are many reading specialists, manybuild a virtual self-portrait. A connector called textbook salespeople, and many principals whoStardoll uses avatars for its online community of at the least can serve as points of reference. Inseven-to-seventeen year-olds, where the members order to assure the registration of new members,can dress and change the outfits of their personal it would seem beneficial to commit to strategiconline doll. Inappropriate and misrepresenting marketing during the initial stages of Web develop-images are largely eliminated. In the case of con- ment – especially if the connector specializes itssumer fraud, eBay is trying to prevent fake listings networking purpose and organizing themes.of expensive consumer goods like plasma televi- There can be a downside to ambitiously seeking out as many new users as quickly as possible, and
  20. 20. Analysis and Evaluation of the Connector Websitethe allure of exponential network growth rates is a customer service meltdown and threaten thetempered by some recent experiences. Friendster Website’s survival. At the outset of Website de-shows how good fortunes can quickly reverse. velopment, it is useful to understand the plannedWith its “six degrees of separation” networking online community’s needs, interests, and priori-technology, Friendster became more popular ties as best as possible. Within these parameters,literally by the week, but technical problems stable connectors prepare and develop the Websitemultiplied as the Website could not keep up with according to a community scale, hoping to serveits growing community of users. Writing in The the projected maximum number of users for thatNew York Times, Gary Rivlin quoted a former online community. Connectors ought to plan forFriendster employee saying: exponential growth rates, but also plan for decel- eration of traffic growth and effectively manageThe service was growing faster than we could user (and investor) expectations.keep up with, so we spent all this time making Fourth, a connector needs to establish a rec-sure the service was stable… A lot of people were ognizable brand, or else it will likely face ficklefrustrated because we weren’t rolling out a lot of users. Three types of branding come to mindfeatures but instead working on infrastructure. for these Websites, either in terms of the online(Rivlin, 2005) community’s social context (i.e. needs, interests, priorities), in terms of community demograph- Another risk for connectors is overestimating ics (i.e. age, gender, race, ethnicity, location),and over-reliance of the viral / word-of-mouth or timing of the Websites launch and entry intomarketing approach. Certainly there are plenty of targeted market(s). Connectors like Classmates.examples of viral messaging by way of connec- com (alumni connections), Match.com (onlinetors, but they are exceptions to what is believed dating), and LinkedIn (professional networking)to be the rule for existing connectors. Pursuing a explicitly focus on the social context of users. Aviral strategy is a fascinating social experiment, social network site like Facebook (teens/college)but it is likely a long-term liability for the host brands itself based on demographics. Sites likeorganization. Sites can cede too much of the eBay and Craigslist have probably benefited aonline community’s development and manage- great deal by the timing of their launches, asment to the discretion and rationale of its users. well as how long they have remained relativelyOrkut is an interesting case in point, having a unchallenged by significant competitors.significant proportion of its user base located in Connectors that do not plan for branding canSouth America. Orkut launched in early 2004 as be criticized for taking the “if you build it, theya challenger to Friendster and MySpace, and it is will come” approach. Friendster seems to be onedoubtful that this Google service initially aimed such example that has not clearly defined a brand.to have the bulk of its online community building The novel applications of social networking,take place in South America. The viral marketing blogging, or tagging, may not be good enough inapproach may be effective for exponential growth the long term to draw new users and to be stickyover a time period, but it can lead to greater long for bringing back users for repeat visits. Theseterm concerns for Website control, development, Websites risk being perceived as overly superficialmanagement, and direction. and faddish, lacking substantive understanding Third, newly launched connectors should of its community of users.seriously consider their scale and benchmarks This may be so, but what about some of thefor sustainable growth. Otherwise, as shown general connectors identified in this report’s databy Friendster, a lack of preparedness may cause analysis that showed rapid growth? For example,
  21. 21. Analysis and Evaluation of the Connector Websiteit is possible that Flickr will establish itself as communications or transactions. User 4.8’s socialthe photo sharing and networking service on capital is actually visible and comes across asthe Internet. Like eBay and Craigslist, Flickr more attractive to others.entered early in the photo-sharing genre, has There are qualitative means for gauginghad few competitors, and built a sizable online trust and reputation on connector Websites.community very quickly. The barriers to market Reviews (also called testimonials, bulletins, andentry are probably considerable for competitors. recommendations) exist on connectors such asHowever, the same could have been said for Facebook, Friendster, MySpace, and Tribe.net.Friendster. It will be interesting to see whether or This user-generated text is an explicit cue. Morenot Flickr fortified market barriers that are higher subtle signals are found on a given user’s profilethan what Friendster had built at its peak before page, such as the appearance of the profile page;the emergence of MySpace and Facebook. Other disclosed frequency of a user’s online activity;connectors like Orkut, Tagged, and Windows quantity and quality of listed friends/contacts;Live Spaces do not appear to neatly fit into any response quality in the profile’s categories; andof the three branding categories. Though all three user postings or comments to message boards.have had tremendous growth since 2005, their Such information can signal personal character-sustainability in the long term may be shakier istics in much the same way we get impressionsthan connectors with recognizable brands. from email. Patricia Wallace has discussed the Fifth, reputation-building mechanisms and impressionistic tendencies that people exhibitaccountability systems likely contribute to traf- online (Wallace, 1999). She explains that givenfic growth. A connector’s legitimacy and value a fixed amount of online information (what coulddepend on how well its feedback systems satisfy be on a connector profile page), people have theuser expectations. Feedback takes many forms, tendency to try to maximize their “impressionranging from discrete to subtle signals. Ratings formation shortcuts” – exerting the least pos-and rankings on connectors like eBay, Facebook, sible amount of energy and work, people will tryand MySpace give explicit cues about other com- to learn as much information as possible aboutmunity members. In effect, these social metric others who are online.systems quantify reputation, and more specifically It seems reasonable that a connector Website’ssocial capital. James Coleman, a very influential online community would like both explicit andsociologist in the late twentieth century, suggested implicit user cues to be as valid and reliable associal capital as “… relations among persons that possible. If this occurs, trust should develop be-facilitate action… it exists in the relations among tween users, and between the community and thepersons.” Ratings and rankings, in particular, Website. If there is little validity and reliability inquantify social capital. They give clear cues, these cues, user flight to a competing connectorrepresenting a given set of relations between two should likely occur over time. Well-constructedusers. They are observable on a connector Web- accountability mechanisms probably temper trendsite, and as a consequence, influence the actions volatility and add to a Website’s stickiness.of other users. Although connectors have had a measurable For example, a user with a 4.8 rating on a 5.0 impact in a variety of social areas (commerce andscale will be viewed as trustworthy and reliable, trading; dating; teenage and collegiate socializing;but someone scoring at 2.2 much less so. Within professional networking), in the future the modelthe social structure consisting of the Website and will likely be used for more specific interests andits community, users have a higher incentive to topics in which people are regularly engagedseek out User 4.8 rather than User 2.2 for online or for people in life transition phases. Themes
  22. 22. Analysis and Evaluation of the Connector Websitecould organize around occupations, education, diaries to find people, information, services, orhealthcare, parenting, religious or spiritual living, other goods.volunteering, politics, hobbies, and residence. In The numbers of people using connectors con-fact, connectors that launched in late 2005 and tinue to swell. Each of at least five (almost six) con-2006 appeared to follow this path toward dif- nectors more than tripled user traffic growth fromferentiation. May 2005 to September 2007. Standout examples Table 7 displays some emerging norms among are Flickr (1,324 %) and LinkedIn (1,117 %). Fivecurrent connector Websites, summarizing the connectors – Flickr, LinkedIn, Orkut, Facebook,main ideas in this section. Thoughtful business and MySpace – showed exponential growth overpractice underscores most of these strategies. the time period. Xanga and LiveJournal wereConnectors add value when identifying com- the only connectors to substantially lose users.munities that either have been underserved in Volatility clearly exists as a characteristic of con-some way or unrecognized by conventional (i.e. nectors, but counter-intuitively, it may be a goodoffline) commercial, government, or nonprofit thing for exponential growth. More research isinstitutions. needed to understand this characteristic, and what it means for connector Websites. It is somewhat speculative here, but connector Websites and theircONcLUsION online communities may need time to mature for attaining sustainability in terms of traffic trends.A new application for Web log analysis is presented Young connector Websites may want to learnin this chapter, and it clearly shows that connector from the coming-of-age experiences of first andWebsites matter a great deal to Americans. This second generation connectors.is evident when observing connectors’ monthly The evidence here is not conclusive, but thetraffic trends and correlating their actual trends Web log analysis does suggest it would be awith an exponential model. They demonstrate mistake to overlook the social implications ofdramatic traffic change over time. Millions of connector Websites. Recent developing storiespeople actively use these Websites as interme- point to future areas for possible research and further examination:Table 7. Emerging Norms for Connector Websites • Understand community in terms of needs, interests, and priorities. • Plan early for community scale. • Clearly set rules of the game to cultivate website norms and values. • Manage user expectations. • Institute accountability systems and trust-building mechanisms. • Structure online incentives to support the website’s norms and values. • Encourage community self-regulation. • Develop infrastructure and capacity as quickly as possible. • Make the website sticky. • Create simple and functional webpages. Optimize user-interface. • Mix content: Content is created by the host, individual users, and host-user collaborations. Features may include user profiles, blogs, wikis, discussion boards, ratings, reviews, rankings, tags, bookmarks, classifieds, chat, interviews, surveys, polls, and web widgets. • Keep pace with fast changing website technologies. • Commit to marketing by word-of-mouth referrals, partnerships, and advertisements. • Synchronize online activities with offline activities. • Establish a brand as quickly as possible.
  23. 23. Analysis and Evaluation of the Connector Website1. Social Values and Tradeoffs: Connector als; family-based networks; pet ownership; Websites offer convenience of choice, reli- the World Cup; mental health issues; youth ability of personal judgment, efficiency of social initiatives; wedding preparations and communications and transactions, and the references; and world travel. Many of these potential for tapping into others’ experi- connectors will never approach the size of ences and resources. What is the downside an eBay, MySpace or Match.com. However to these expanded freedoms? Issues of they are likely establishing a core compe- personal privacy and security have hit the tency and competitive advantage based on headlines in 2006. Stories grow about pedo- substance, and not relying solely on the philes stalking on teen connector Websites novelty of their social networking applica- like MySpace ready to prey on vulnerable tion, or the timing of their Website launch. or overly trusting young girls and boys. A Increased specialization should continue public reaction was inevitable. The U.S. in the future, and the community scale of House Committee on Energy and Commerce connector Websites on average will probably Subcommittee on Oversight and Investiga- shrink as the overall sector matures. tion held several hearings from June 21-28, 2006, which included testimony from the Connector Websites continue to adapt to the Federal Trade Commissioner, federal and prevailing social and economic circumstances state law enforcement officials, and ex- of the day. They have been at the forefront of the ecutives from Facebook, MySpace, Xanga, so-called Web 2.0 online expansion era. In terms Google, Yahoo!, Microsoft, and market of participation numbers, Website traffic trends, research organizations. Major concerns and unorthodox business practices and strategies, aired about the confidentiality of members’ connector Websites are a timely topic. personal data and how Websites monitor and The overarching goal here has been to bridge enforce safety, rules, online community best the somewhat disconnected information pools practices, and adherence to the Children’s and audiences coming out of academia, the main- Online Privacy Protection Act (COPPA). stream news media, bloggers, and the Internet The Federal Trade Commission now has a industry’s insiders and analysts. Other objectives “Facts for Consumers” bulletin for parents included: (1) Propose and define the connector posted on the agency Website. It is unclear Website as a new type of social institution and which direction public policy will turn at intermediary; (2) Describe how the connector this point, but it is possible state and federal Website model is playing out in the real world; government officials may seek to directly (3) Apply an institutional approach to analyze the regulate connector Websites. behavior of connector Websites and their online2. Differentiation and Specialization: As communities; (4) Determine whether or not con- connector Websites evolve, they appear to be nector Websites have demonstrated exponential differentiating with respect to their organiza- change over time; and (5) Suggest the implications tional missions and focusing branding efforts of existing connector Websites, pointing to future on users’ needs, interests, and priorities. This trends and possible areas for research. The use started to happen in 2006. A sampling of of Web log analysis in this study should enable new connector themes and organizing top- us to better evaluate connector Websites as they ics are related to: political campaigns and evolve in the future. elections; religious expression and sharing; car and truck enthusiasts; female profession-
  24. 24. Analysis and Evaluation of the Connector WebsiteAcKNOWLEDGMENt Boase, J., Horrigan, J., Wellman, B., Rainie, L. (2006). The strength of Internet ties. Washington,I would like to thank the following people for DC: Pew Internet American Life Project.helpful discussions, reviews, and suggestions over Boyd, D. (2004). Friendster and publicly articu-the past two years: Joe Carney, Jim Cashel, Joe lated social networks. Proceedings of ACM Con-Cothrel, Sunil Dasgupta, Jim DiPerna, Matt Dull, ference on Human Factors in Computing Systems,Katie Field-Mateer, Patrick Gavin, Joe Geraghty, New York: ACM Press, 1279-1282.Daniel Harrison, Jim Harvey, Eszter Hargittai,Jim Jansen, Ben Klemens, Katharine Kravetz, Boyd, D. (2006). Friends, Friendsters, andAndrew Lee, Amanda Lenhart, Kate Mazukelli, MySpace Top 8: Writing community into beingSarah-Kay McDonald, Anjetta McQueen, Elana on social network sites. First Monday, 11(12).Mintz, Alan Murphy, Erin Murphy, Jason Palmer, Boyd, D., Heer, J. (2006). Profiles as con-Jonathan Rauch, Stephen Rose, Judy Smith-Davis, versation: Networked identity performance onFred Stutzman, Russell Wheeler, Dave Witzel, Friendster. Proceedings of Thirty-Ninth Hawai’iDanny Yagan, and participants at the 2005 On- International Conference on System Sciences.line Community Summit in Sonoma, CA. Their Los Alamitos, CA: IEEE Press.unique perspectives and experiences have beeninvaluable resources for writing this report. Boyd, D., Ellison, N. (2007). Social network A big thank you to Andrew Lipsman and sites: Definition, history, and scholarship. JournalcomScore Media Metrix for providing me with of Computer-Mediated Communication, 13(1),the website traffic data used in my analysis. com- article 11.Score Media Metrix is a leading website tracking Byrne, D. N. (2007). Public discourse, communityorganization that is based in Reston, Virginia. concerns, and civic engagement: Exploring black Finally, words cannot express how grateful I social networking traditions on BlackPlanet.com.am to my wife Christy for her support, honesty, Journal of Computer-Mediated Communication,and comments throughout this entire research 13(1), article 16.project. Coase, R. (1998). The new institutional economics. The American Economic Review, 88(2), 72-74.rEFErENcEs Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of Sociology:Adamic, L. A., Buyukkokten, O., Adar, E. Supplement: Organizations and Institutions:(2003). A social network caught in the Web. First Sociological and Economic Approaches to theMonday, 8(6). Analysis of Social Structure, 94, S100-S101.Backstrom, L., Huttenlocher, D., Kleinberg, J., Coleman, J. S. (1990). Foundations of socialLan, X. (2006). Group formation in large social theory. Cambridge: Harvard University Press,networks: Membership, growth, and evolution. 300-316.Proceedings from the Annual InternationalConference on Knowledge Discovery and Data DiPerna, P. (2006). K-12 encounters the Internet.Mining (Association for Computing Machinery First Monday, 11(5).KDD). Philadelphia, Pennsylvania. Dodds, P., Muhamad, R., Watts, D. (2003).Barnes, S. (2006). A privacy paradox: Social An experimental study of search in global socialnetworking in the United States. First Monday, networks. Science, 301, 827-829.11(9).