ChapterPDF Available

The Impact of Website Design on Users’ Trust Perceptions


Abstract and Figures

With rapid digitalization, trust has become a critical issue in designing and maintaining e-commerce platforms – without trust, no transaction takes place. Companies that design for trust have a strategic advantage over competitors. Although trust is a crucial factor in e-commerce, designing a trustworthy website can be challenging for companies that make most or part of their profits online. The study builds on prior research to propose a comprehensive and up-do-date checklist. Trust components are divided into three dimensions of website design; their impact on users’ trust perceptions is studied in an online experiment with two websites. The results present demonstrable evidence that website design has a powerful impact on users’ trust perceptions. Professional design, primarily visual aspects, is responsible for creating a positive user first impression. Furthermore, additional trust is built through different dimensions of website design, increasing the likeliness of buying from the site.
Content may be subject to copyright.
The Impact of Website Design on Users’ Trust
Kairi Fimberg, Sonia Sousa
Tallinn University, School of Digital Technologies,
25 Narva Rd, 10120 Tallinn, Estonia,
Abstract. With rapid digitalization, trust has become a critical issue in design-
ing and maintaining e-commerce platforms – without trust, no transaction takes
place. Companies that design for trust have a strategic advantage over competi-
tors. Although trust is a crucial factor in e-commerce, designing a trustworthy
website can be challenging for companies that make most or part of their profits
online. The study builds on prior research to propose a comprehensive and up-
do-date checklist. Trust components are divided into three dimensions of web-
site design; their impact on users’ trust perceptions is studied in an online ex-
periment with two websites. The results present demonstrable evidence that
website design has a powerful impact on users’ trust perceptions. Professional
design, primarily visual aspects, is responsible for creating a positive user first
impression. Furthermore, additional trust is built through different dimensions
of website design, increasing the likeliness of buying from the site.
Keywords: Website Design · User Trust · Trust Components · Trust Design ·
Visual Design · Content Design · Social-Cue Design · Trust Assessment
1 Introduction
A good part of communication previously carried out between humans now relies on
technology and human-computer interaction. With Internet technologies and infra-
structures to support e-commerce mostly established although always evolving ,
focusing on the psychological factors that affect e-commerce acceptance by online
users has been on the rise. One such factor, playing a significant role in the success of
e-commerce, is trust [17].
For a company, a website is often the first point of contact for potential customers,
responsible for first impressions and generating revenue. Companies that design for
trust have a strategic advantage over competitors. Without trust, visitors leave the
website immediately. The probability of leaving is higher during the first seconds.
While most companies know they need to have trust components on a website, they
are often overlooked because sometimes, it does not provide measurable value [3].
Research problem and significance. It has been shown that web browsing exhib-
its a significant “negative aging” phenomenon, meaning that some initial screening
has to be passed before a page is examined in detail, giving rise to the browsing be-
11th International Conference on Applied Human Factors and Ergonomics (AHFE 2020). Springer,. (Human Factors in
Communication of Design) [in print]
havior called “screen-and-glean” [12]. To gain only several minutes of users’ atten-
tion, a website must clearly communicate its value proposition within 10 seconds
[13]. A study by Bentley University Design and Usability Center demonstrates that
when people first navigate to a website, businesses have about 6 seconds to create a
positive impression with users [1]. This is an extremely small window of time to con-
vince users that the website is one that they can trust making a purchase from.
Although trust is a crucial factor in e-commerce and the concept has been widely
studied by several authors, a comprehensive, easy-to-follow checklist of components
that companies could utilize when designing their website was either missing or out-
dated. Nevertheless, almost every company struggles with website credibility, as visi-
tors are immediately skeptical [14].
Goal of the research. In line with the aforementioned shortcoming, the study aims
to understand the components in website design needed to build a trustworthy web-
site; and to propose an up-do-date design checklist. The following research questions
were formulated to reach this goal.
RQ1: How important is visual design as an initiator of trust?
RQ2: How is trust influenced by different dimensions of website design?
RQ3: How important is trust-inducing design for the purchase decision?
2 Background
Trust is a common and essential concept in different domains, the term has been de-
fined in different ways, and there is no widely-accepted definition [15, 6]. Trust and
trust relationships in the offline world have been a topic of research in various disci-
plines, such as philosophy, sociology, psychology, management, marketing, etc.; and
each of these disciplines has produced its own concepts, definitions, and findings [7].
As a social being, trust is incorporated into every aspect of human life. Trust is es-
sential for all kinds of personal relationships, “the loom on which is woven the social
fabric of society” [4]. Without trust, social life breaks down, no business transactions
take place, new technology is not adopted, and even political legitimacy collapses
[11]. Trust is the basis for decision making in many contexts, and the motivation for
maintaining long-term relationships based on cooperation and collaboration [6].
As emphasized by several authors, trust is the key to the success of e-commerce
[16, 17, 11], and a prerequisite for actions involving another agent in which one may
suffer physical, financial, or psychological harm [2]. Lack of trust has been identified
as one of the most formidable barriers for engaging in e-commerce, involving transac-
tions in which financial and personal information is shared [8, 18, 19].
Establishing trust relationships in a digital environment involves more aspects than
in the social world because communications in the computing network rely on not
only relevant human beings and their relationships but also digital components [19].
The model of trust for e-commerce (MoTEC) by Egger provides a framework of
six components, regrouped into three more significant categories: pre-interactional
filters (pre-purchase knowledge), interface properties, and informational content [8].
To render the design process more coherent, the model components have been redis-
tributed into three qualitatively different types of requirements: appeal (graphic de-
sign), usability (structure and navigation design), and trustworthiness (content de-
sign). Cheskin Research focuses on website interface cues and presents a model of six
primary components that play a major role in communicating trustworthiness [5]. The
building blocks of trust are seals of approval, brand, technological sophistication,
navigation, presentation, and fulfillment. These blocks, in turn, can be divided into a
total of 28 components that can be used to communicate functional trustworthiness.
Patel lists over 40 factors that influence website credibility [14], using the four types
of credibility by Fogg and Tseng [9]. Presumed credibility: general assumptions in the
perceiver’s mind; reputed credibility: what third parties have reported; surface credi-
bility: based on simple inspection; earned (experienced) credibility: first-hand experi-
ence, reputation built over time. According to Patel, “The goal with each of these
credibility factors is to stack the deck in your favor” [14].
From the perspective of the current study, grouping trust components within web-
site design dimensions seems to be the most fitting framework. Elaborating on the
framework by Wang [17], that allocates trust components in three design dimensions
(visual, content, and social-cue design), and considering the components of a trust-
worthy website suggested by other cited authors, the author of the current study pro-
poses a refined checklist of trust-inducing design components (shown in Appendix).
The last dimension, the social-cue design, is amended to incorporate both social pres-
ence and social proof components.
3 The Study
To answer the research questions and test the checklist of trust components proposed,
an experiment was carried out online, focusing on the quantitative results collected
from two different websites of Estonian furniture manufacturers. The A/B test fea-
tured a between-subjects study design. The questionnaire guided participants through
four separate parts: first impression, design assessment, trust assessment, and final
comments. A 7-point Likert scale was used for all questions. To evaluate participants’
trust perception, the trust assessment model by Gulati et al. [10] was put into practice., an online user experience and screen recording platform, was used to
gain additional insight into users’ browsing behavior.
A pilot study was carried out on a small group of participants to evaluate the time
and statistical effect to predict the appropriate sample size, plus test the experiment
protocol. The sampling technique used was convenience sampling, which is often
used in business studies to gain initial primary data regarding specific issues like the
perception of an image of a particular brand or opinions of a new design.
A total of 50 participants were recruited for the study. Participants were randomly
assigned into two groups, 25 in each. In Group A, there were 17 female and 8 male
participants. The majority of them fell in two age groups, 25-35 (13 participants) and
35-44 (11 participants); 1 participant was older, aged 55-64. In Group B, the gender
was slightly more equal; there were 14 female and 11 male participants. The majority
of them fell in the same two age groups, 25-35 (9 participants) and 35-44 (15 partici-
pants); again, 1 participant was older, aged 55-64. Most of the participants in Group A
and B were Estonians, with 2 and 1 Russian respectively. In both groups, 23 out of 25
participants had higher (tertiary) education, 2 had secondary education. A majority of
participants shopped online regularly.
4 Results and Discussion
The results of the study present demonstrable evidence that website design has a pow-
erful impact on users’ trust perceptions. The data revealed that the website with an
attractive and contemporary design (the perceived average quality of design in Web-
site A was 5.71; first impression 5.67) implicated considerably higher trust than the
website with a dull and outdated design (the perceived average quality of design in
Website B was 3.54; first impression 3.16).
Fig. 1. Perceived quality of website design (left) and perceived trust (right) in Website A (blue)
and Website B (red). The three design dimensions are visual design (VD), content design (CD),
and social-cue design (SD). Altogether 21 trust-inducing design components were evaluated.
In order to measure users’ trust perceptions, participants were asked about benevo-
lence (BEN), competency (COMP), reciprocity (REC), risk (RISK), and general trust
(GEN) – the constructs coined by Gulati et al. [10]. The overall trust level of Website
A was 6.05 (86.41%), while the overall trust level of Website B was 3.88 (55.39%).
Based on the trust model, scores between 80-90% are considered as a high trust level,
while scores between 50-60% are considered as a very low trust level.
To find answers to the research questions, correlation analysis was done (see Table 1).
RQ1 strong positive correlation. Participants’ browsing behavior confirmed
that when visitors first come to a website, they develop their first impression in a mat-
ter of seconds. The average time spent on Website A was 1 minute 16 seconds, during
which, on average, 7.81 clicks were made by 87.5% of participants. In contrast, the
average time spent on Website B was only 32 seconds, during which, on average, 2.17
clicks were made by 25% of participants.
Professional design, primarily visual aspects, is responsible for creating a positive
first impression, which, in turn, is strongly correlated with trust, leading us to believe
that visual design is a vital initiator of trust.
RQ2 strong or moderate positive correlation (with some exceptions). The in-
fluence on users’ trust perceptions is active within all three design dimensions (visual,
content, and social-cue design), indicating the importance of them all when creating
or redesigning a website. There was no significant correlation between some of the
design components and trust. This, however, does not mean that website visitors think
of these components as unimportant. It tells us that trust perceptions are not based on
one component only but rather on a collection of them. Lacking in quality of some
components does not significantly decrease the overall trust level, as having a few key
components does not increase users’ trust to a sufficient level.
RQ3 strong positive correlation. Similarly, the study affirmed the importance
of trust-inducing website design for the visitor’s purchase decision, i.e., whether the
website succeeds in converting a visitor into a customer. The process chain here is the
following: strong design fosters higher trust; higher trust makes it more likely that a
visitor engages in a purchase decision. However, as with the first impression and in-
terest, visitors’ purchase decisions also depend on other factors, like their actual need
for the product or service, and whether they can afford it.
Table 1. Correlation analysis.
Correlations (Pearson)
of buying
First impression
Visual design
Correlations (Pearson)
VD1 [Design]
CD1 [Brand information]
VD2 [Color]
CD2 [Company information]
VD3 [Font]
CD3 [Contact information]
VD4 [Images]
CD4 [Content]
VD5 [Search]
CD5 [Blog / news page]
VD6 [Navigation]
CD6 [Grammar]
VD7 [Links, buttons, forms]
CD7 [Product information]
VD8 [Technical functioning]
CD8 [Price information]
CD9 [Policies]
SD1 [Customer service]
CD10 [Guarantees / warranties]
SD2 [Social presence]
CD11 [Important questions]
** Correlation is significant at the 0.01 level (2-tailed). N = 50.
VD = visual design, CD = content design, SD = social-cue design.
The independent samples t-test was used to determine whether there was statistical
evidence that the means of two groups were significantly different. All results (first
impression, design, trust, likeliness of buying) came back as significantly different.
The trust assessment model by Gulati et al. [10] proved to be a reliable tool for
measuring trust in a website, with the Cronbach’s alpha coefficient of 0.969.
5 Limitations and Further Research
The main limitation was the stimuli. The two websites used in the study differed in
terms of product design and style. The product itself can influence the first impres-
sion, trust, and the likeliness of buying. Although the results of Website A were sig-
nificantly different from those of Website B, ideally, we should have an identical
product that is unknown to all participants, with the only difference being the website
design, if we wanted to measure solely the impact of website design on users’ trust.
While the setup of this study was able to list the trust-inducing components and
provide quantitative results to confirm the relationship between design and trust, fur-
ther research is required to investigate the importance of these components in differ-
ent situations (their effects based on users’ gender, age, and a company’s field of
business), and how to strategically place them into the website.
A more detailed discussion and suggestions can be found in the author’s full thesis.
1. Albert, B. (2012, June 6): How Quick Are We to Judge? A Case Study of Trust and Web
Site Design. Retrieved January 9, 2019,
2. Bickmore, T., Cassell, J.: Relational Agents: A Model and Implementation of Building
User Trust. In: Proceedings of the SIGCHI Conference on Human Factors in Computing
Systems, 3(1) (2001)
3. Blue Fountain Media (n.d.): 8 Trust Factors to Help Build a Website and Convert Users
[Blog post]. Retrieved January 10, 2019,
4. Cassell, J., Bickmore, T.: External Manifestations of Trustworthiness in the Interface. In:
Communications of the ACM, 43(12) (2000)
5. Cheskin Research (2000, July): Trust in the Wired Americas. Retrieved January 15, 2019,
6. Cho, J.-H., Chan, K., Adali, S.: A Survey on Trust Modeling. In: ACM Computing Sur-
veys (CSUR), 48(2) (2015)
7. Corritore, C.L., Kracher, B., Wiedenbeck, S.: On-line Trust: Concepts, Evolving Themes,
a Model. In: International Journal of Human-Computer Studies, 58(6) (2003)
8. Egger, F.N.: Trust Me, I’m an Online Vendor: Towards a Model of Trust for E-Commerce
System Design. In: Extended Abstracts on Human Factors in Computing Systems (2000)
9. Fogg, B.J., Tseng, H.: The Elements of Computer Credibility. In Proceedings of the
SIGCHI Conference on Human Factors in Computing Systems (1999)
10. Gulati, S., Sousa, S., Lamas, D.: Modelling Trust in Human-like Technologies. In: Pro-
ceedings of the 9th Indian Conference on Human Computer Interaction (2018)
11. Hasan, Z., Gope, R.C., Uddin, M.N.: Do Aesthetics Matter in Long-Established Trust? In:
International Journal of Computer Applications, 77(13) (2013)
12. Liu, C., White, R.W., Dumais, S.: Understanding Web Browsing Behaviors through
Weibull Analysis of Dwell Time. In: Proceedings of the 33rd International ACM SIGIR
Conference on Research and Development in Information Retrieval (2010)
13. Nielsen, J. (2011, September 12): How Long Do Users Stay on Web Pages? Retrieved
January 15, 2019,
14. Patel, N. (n.d.): 41 Factors That Influence Your Website’s Credibility [Blog post]. Re-
trieved January 15, 2019,
15. Rousseau, D.M., Sitkin, S.B., Burt, R.S., Camerer, C.: Not So Different After All: A
Cross-discipline View of Trust. In: The Academy of Management Review, 23(3) (1998)
16. Tan, F.B., Sutherland, P.: Online Consumer Trust: A Multi-Dimensional Model. In: Jour-
nal of Electronic Commerce in Organizations (JECO), 2(3) (2004)
17. Wang, Y.D.: Trust in B2C E-Commerce Interface. In: Encyclopedia of Information Sci-
ence and Technology (1st ed.) (2005)
18. Wang, Y.D., Emurian, H.H.: An Overview of Online Trust: Concepts, Elements, and Im-
plications. In: Computers in Human Behavior, 21(1) (2005)
19. Yan, Z., Holtmanns, S.: Trust Modeling and Management: from Social Trust to Digital
Trust. In: Computer Security, Privacy and Politics: Current Issues, Challenges and Solu-
tions (2008)
Appendix: The Design Checklist
Trust-inducing design components
Defines the graphical
design aspect and the
structural organization of
displayed information on
the website.
ü Professional design
ü Color scheme to suit the product/service
ü Nice and legible fonts
ü High-quality (and authentic) images and
ü Good on-site search
ü Easy-to-use navigation
ü Clear anchor text and microcopy
ü No technical problems (broken links,
missing pictures or pages)
Refers to the informa-
tional components that
can be included in the
website, be they textual,
graphical, etc.
ü Brand-promoting information (logo,
ü Company information (“About” page,
facts & figures)
ü Contact information
ü Physical address
ü Useful (expert-level) content
ü Good grammar, minimized jargon
ü External links (sources)
ü Up-to-date blog / news page
ü Clients (client logos)
ü Client case studies
ü Product information
ü Price information
ü Order information (transaction reports)
ü Clear policies (privacy, return)
ü Guarantees and warranties
ü Helpful FAQs
ü Trust seals
ü Awards
Relates to embedding
social and interpersonal
cues, such as social
proof, social presence
and face-to-face interac-
tion, into the website via
different (communica-
tion) media.
ü Staff photos and bios
ü Easy access to customer service (e.g.,
contact form)
ü Instant messaging / chat option
ü Social presence (social media)
ü Testimonials
ü Reviews
ü Reviews from influencers and notable
ü Professional product reviews (from re-
view sites, bloggers, customers)
ü Press articles (media logos)
... With this use of dark design patterns to persuade or manipulate user experience, it proliferates conversations about ethical design (Waldman 2020) and the need for trust in technology (Oper and Sousa 2020;Fimberg and Sousa 2020), highlighting as well the effects of fake news in cognitive bias (Shu et al. 2017) in assigning trust in the information source (Trust 2018; Gulati et al. 2019). ...
Full-text available
This study reports an exploratory inquiry into the problematic phenomenon of fake news on Facebook, aiming at providing an inside view on how users in the United Kingdom (UK) value the credibility of news posts on Facebook in a post-Brexit era. Participants ( n = 201) were asked to review four different Brexit-related Facebook posts that linked to news articles from UK tabloids that were published between 2016 and 2019. Two of the posts were debunked as fake news, while the other two were verified as real news. The authors of each Facebook post were different: two from UK tabloids and two from unknown individuals. Respondents were asked to identify the credibility of the news posts in Facebook’s news feed. The results indicate that the author of the post significantly influences users’ perceived credibility. For instance, a fake news post from an individual is perceived as the least trustworthy, while a real news post from an individual and a fake news post from a tabloid are somewhat similarly perceived. The content of a post is seen as most trustworthy when it is a real news post from a tabloid and as least credible when it is a fake news post from an individual. Finally, in two cases, credibility can predict willingness to interact with a post. The research concludes with a set of recommendations for future research.
It is expected that website design can lead to online trust, although research that systematically examines such a relationship is little. In addition, far too little attention has been paid to understand the effect of website design on e-trust across gender. In this study, therefore, three elements of website design (visual design, information design, and navigation design) are examined for their effect on consumer trust in e-commerce. Using data collected from 532 online shoppers , the overall model was first tested using structural equation modeling (SEM) analysis. The research model was then tested for each gender group separately. The results of the study indicated that website design is an important factor in establishing trust in e-commerce. Our results further revealed that website information design is more important to males than females in forming e-trust, while website navigation design is more important to females than males in forming e-trust. However, website visual design was found to be a key driver of e-trust for both males and females.
This chapter addresses how the pandemic has transformed the world at a surprising speed in terms of consumer behaviors and applicable virtual strategies to capture new opportunities. Although there have been some short-term sharp declines in consumer spending, confident changes have been realized through consumers' evolution into “digitalization” or the use of contactless, functional, and appropriate business models. Accordingly, the pandemic has also brought digitalization forward for companies, making it almost obligatory. This chapter will also discuss exemplary, inspirational practices of companies that have developed business models, products, and services by managing changes accelerated by the pandemic.
Conference Paper
Full-text available
Trust is an important decision-making construct helping users to adopt and continually use a system. Despite its importance, past research has mainly focused on studying and researching it in technology mediated interactions. As nature of technical artifacts becomes increasingly intelligent and human-like, a theory for explaining and predicting trust in such artifacts also becomes important. However, studying trust relationship of a user with technical artifacts is still in its infancy. To address this, we build on prior work which empirically modelled trust in user technology interactions and evaluate a human computer trust model using partial-least square structural equation modelling (PLS-SEM). Our research results identify four factors: competence, benevolence, willingness & reciprocity which impact user trust. Suggestions for future applications of the model are also discussed. Our study contributes to the literature by advancing the discussion on (a) human-like technologies and (b) the role of trust in human-artifact relationship.
Full-text available
The concept of trust and/or trust management has received considerable attention in engineering research communities as trust is perceived as the basis for decision making in many contexts and the motivation for maintaining long-term relationships based on cooperation and collaboration. Even if substantial research effort has been dedicated to addressing trust-based mechanisms or trust metrics (or computation) in diverse contexts, prior work has not clearly solved the issue of how to model and quantify trust with sufficient detail and context-based adequateness. The issue of trust quantification has become more complicated as we have the need to derive trust from complex, composite networks that may involve four distinct layers of communication protocols, information exchange, social interactions, and cognitive motivations. In addition, the diverse application domains require different aspects of trust for decision making such as emotional, logical, and relational trust. This survey aims to outline the foundations of trust models for applications in these contexts in terms of the concept of trust, trust assessment, trust constructs, trust scales, trust properties, trust formulation, and applications of trust. We discuss how different components of trust can be mapped to different layers of a complex, composite network; applicability of trust metrics and models; research challenges; and future work directions.
Full-text available
Recently trust has gained momentum in user experience (UX) research due to the rapid development of e-commerce services, privacy and security issues. It is widely considered that the perception of trust is not only influenced by the website attributes (i.e., usability, functionality), but also by other human factors such as aesthetics. However, it is not clear that which factors of aesthetics influence what dimensions of trust while considering aesthetics and trust as multidimensional constructs. In this paper, we present insights from an experiment conducted with 30 participants on an e- commerce website displayed on an 15 inch display. Participants' responses on trust, aesthetics and perceived usability and user satisfaction were collected during the experiment.
Full-text available
Lack of trust has been repeatedly identified as one of the most formidable barriers to people for engaging in e-commerce, involving transactions in which financial and personal information is submitted to merchants via the Internet. The future of e-commerce is tenuous without a general climate of online trust. Building consumer trust on the Internet presents a challenge for online merchants and is a research topic of increasing interest and importance. This paper provides an overview of the nature and concepts of trust from multi-disciplinary perspectives, and it reviews relevant studies that investigate the elements of online trust. Also, a framework of trust-inducing interface design features articulated from the existing literature is presented. The design features were classified into four dimensions, namely (1) graphic design, (2) structure design, (3) content design, and (4) social-cue design. By applying the design features identified within this framework to e-commerce web site interfaces, online merchants might then anticipate fostering optimal levels of trust in their customers.
Full-text available
Consumer trust is widely heralded as an essential element for the success of electronic commerce, yet the concept itself is still clouded in confusion by the numerous definitions and points of view held across multiple disciplines. It is the aim of this paper to synthesise the viewpoints from across the disciplines and bring them together in a multi-dimensional trust model. It is hoped that in doing so, these broad assortment of views will highlight the true underlying nature of trust in the online environment. From these multiple disciplines, three dimensions of trust emerged: dispositional trust, institutional trust and interpersonal trust, each bringing its own influencing factors into the overall intention to trust. From this model emerged the notion that the consumer as an individual is central to the understanding of trust, and in turn that the individual’s personality and culture form the foundation for the development of trust.
This chapter introduces trust modeling and trust management as a means of managing trust in digital systems. Transforming from a social concept of trust to a digital concept, trust modeling and management help in designing and implementing a trustworthy digital system, especially in emerging distributed systems. Furthermore, the authors hope that understanding the current challenges, solutions and their limitations of trust modeling and management will not only inform researchers of a better design for establishing a trustworthy system, but also assist in the understanding of the intricate concept of trust in a digital environment.
Trust is emerging as a key element of success in the on-line environment. Although considerable research on trust in the offline world has been performed, to date empirical study of on-line trust has been limited. This paper examines on-line trust, specifically trust between people and informational or transactional websites. It begins by analysing the definitions of trust in previous offline and on-line research. The relevant dimensions of trust for an on-line context are identified, and a definition of trust between people and informational or transactional websites is presented. We then turn to an examination of the causes of on-line trust. Relevant findings in the human–computer interaction literature are identified. A model of on-line trust between users and websites is presented. The model identifies three perceptual factors that impact on-line trust: perception of credibility, ease of use and risk. The model is discussed in detail and suggestions for future applications of the model are presented.
Conference Paper
Given the importance of credibility in computing products, the research on computer credibility is relatively small. To enhance knowledge about computers and credibility, we define key terms relating to computer credibility, synthesize the literature in this domain, and propose three new conceptual frameworks for better understanding the elements of computer credibility. To promote further research, we then offer two perspectives on what computer users evaluate when assessing credibility. We conclude by presenting a set of credibility-related terms that can serve in future research and evaluation endeavors.
Conference Paper
Dwell time on Web pages has been extensively used for vari- ous information retrieval tasks. However, some basic yet im- portant questions have not been sufficiently addressed, e.g., what distribution is appropriate to model the distribution of dwell times on a Web page, and furthermore, what the dis- tribution tells us about the underlying browsing behaviors. In this paper, we draw an analogy between abandoning a page during Web browsing and a system failure in reliabil- ity analysis, and propose to model the dwell time using the Weibull distribution. Using this distribution provides bet- ter goodness-of-fit to real world data, and it uncovers some interesting patterns of user browsing behaviors not previ- ously reported. For example, our analysis reveals that Web browsing in general exhibits a significant "negative aging" phenomenon, which means that some initial screening has to be passed before a page is examined in detail, giving rise to the browsing behavior that we call"screen-and-glean." In addition, we demonstrate that dwell time distributions can be reasonably predicted purely based on low-level page fea- tures, which broadens the possible applications of this study to situations where log data may be unavailable. Categories and Subject Descriptors