Location and time do matter: A long tail study of website requests Chetan Kumar a, ⁎, John B. Norris b , Yi Sun a a Department of Information Systems and Operations Management, College of Business Administration, California State University San Marcos, 333 South Twin Oaks Valley Road, San Marcos, CA 92096, United States b Krannert School of Management, Purdue University, 403 West State Street, West Lafayette, IN 47907, United States abstract article info Article history: Received 16 October 2008 Received in revised form 15 April 2009 Accepted 17 April 2009 Available online 3 May 2009 Keywords: Website visitations Long tail model Request heterogeneity User location Time of access There has been a tremendous growth in the amount and range of information available on the Internet. The users' requests for online information can be captured by a long tail model. A few popular websites enjoy a high number of visitations while the majority of the rest are less frequently requested. In this study we use real world data to investigate this phenomenon and show that both users' physical location and time of access affect the heterogeneity of website requests. The effect can partially be explained by differences in demographic characteristics at locations and diverse user browsing behavior in weekdays and weekends. These results can be used to design better online marketing strategies, affiliate advertising models, and Internet caching algorithms with sensitivities to user location and time of access differences. © 2009 Elsevier B.V. All rights reserved. 1. Introduction There has been a tremendous growth in the amount and range of information available on the Internet. Cisco Systems report [10] forecasts that global Internet Protocol (IP) traffic will increase six folds between 2007 and 2012, from fewer than 7 exabytes per month in 2007 to 44 exabytes per month in 2012. ComScore Inc. [12] estimates that total global Internet users have surpassed 1 billion visitors in December 2008. The increase in Internet traffic is aided because making information available online is becoming relatively inexpen- sive, and as more people have Internet access demand for information increases. In addition, new format and content of Web 2.0 technol- ogies such as video, social networking and collaboration applications prompted more interest in online deliveries. The trend of increasing Internet traffic is likely to continue [10,13]. The visitation of users to websites or online product purchase can be captured by a long tail model coined by Chris Anderson [2], shown in Fig. 1 . Anderson [2] used this model to explain the success of Amazon book and Netflix DVD rental recommendations system to promote obscure products. Brynjolfsson et al. [7] had earlier noted this effect due to lower search costs in the digital economy. A few popular websites enjoy a high number of visitations. Interestingly there are also a large number of infrequently requested websites. The former is shown by the steep end of the curve, while the latter forms the tapering long tail. Before the Internet age economic scale favored services that catered to a large amount of customers. For example, books that potentially attract more readers will be more likely published than those targeting niche markets [2]. However the inexpensive online medium and reduced intermediaries lowered the hurdle of entrance. Websites with potentially small audiences can also exist because of relatively inexpensive hosting costs for the information service provider. All sorts of information has a more equal chance to be present on the Internet and, with the assistance of efficient search engines such as Google, to be found by users. In this study we investigate this phenomenon by using real world data and show how users' location and time of access (weekdays versus weekends) affects this long tail model. Our results can be used to design better online marketing strategies, affiliate advertising models, and Internet caching algo- rithms. According to Interactive Advertising Bureau Internet ad revenues grew to $21 billion in 2007, up 25% over 2006. However as online advertising is still only 10% of all US ad spending it has considerable room to grow [3]. Internet based marketing, advertising, and content delivery is becoming increasingly important for busi- nesses and institutions. Understanding patterns in user request behavior can provide guidelines for customizing advertising pricing for Internet portals such as Yahoo and Google. In addition they can be exploited for Internet caching algorithms to reduce user delays on the Internet. Therefore we believe this study contributes to an important area for Information Systems research. The rest of this paper is organized as follows. We first discuss literature related to online visitation and marketing, followed by a description of methodology of analysis. Next we present the model Decision Support Systems 47 (2009) 500–507 ⁎ Corresponding author. Tel.: +1760 477 3976; fax: +1 760 750 4250. E-mail addresses: ckumar@csusm.edu (C. Kumar), johnbnorris@alumni.purdue.edu (J.B. Norris), ysun@csusm.edu (Y. Sun). 0167-9236/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.dss.2009.04.015 Contents lists available at ScienceDirect Decision Support Systems journal homepage: www.elsevier.com/locate/dss