We present results from a qualitative study examining how professionals living and working in Nairobi, Kenya regularly use ICT in their everyday lives. There are two contributions of this work for the HCI community. First, we provide empirical evidence demonstrating constraints our participants encountered when using technology in an infrastructure-poor setting. These constraints are limited bandwidth, high costs, differing perceptions of responsiveness, and threats to physical and virtual security. Second, we use our findings to critically evaluate the “access, anytime and anywhere” construct shaping the design of future technologies. We present an alternative vision called deliberate interactions—a planned and purposeful interaction style that involves offline preparation—and discuss ways ICT can support this online usage behavior.
Web & Social Media Analytics Previous Year Question Paper.pdf
Chi.talk
1. Deliberate Interactions : Characterizing Technology Use in Nairobi, Kenya Susan P. Wyche Thomas N. Smyth Marshini Chetty Paul M. Aoki (Intel Labs Berkeley) Rebecca E. Grinter
13. Finding: High Costs High costs inhibited Internet access at home. “ If I wanted Internet access I could get it, but I go to work every morning and we have access. It cost about 12,000 Shilling [USD $157] for the whole year. I don’t think I need to spend that much, especially when I can access it the next morning at work.” Kenyan Schillings
15. Findings: High Costs “ I think the flash disk is one of the best things to come out, because before, I used to save my research on a floppy disk, the capacity was very small and by the time I saved one or two web pages it’s full, so my flash disk helps me to download much of the info I want, then I go home to analyze them .” Our participants made a point of using online time at cafes as efficiently as possible.
16.
17. Finding: Responsiveness Designers of collaborative work systems typically assume that communication occurs between individuals and groups with similar Internet speeds and connectivity.
18. “ This week, I was working at home . . . and I missed email. I had 174 messages, 174 messages! From South Africa, from Washington D.C., and I have to reply and send things. I don’t have Internet at home, I can not respond to emails as quickly as they can.” Finding: Responsiveness
19. Finding: Responsiveness Managing different expectations for responsiveness and alerting those in more connected environments about their colleagues’ intermittent and (at best) slow Internet access was a constant struggle for our participants.
20. Finding: Responsiveness Some participants told us they were uninterested in adopting a work culture that relied so heavily on fast-paced email exchanges, rather than face-to-face communication. Discussions about the “cultural bias” embedded in ICT have been presented elsewhere and reported when ICT were first appearing in Kenya (Van Ryckeghem, 2002).
21.
22. Contributions 1) We provide empirical evidence demonstrating constraints our participants encountered when using technology in an infrastructure-poor setting (limited bandwidth, high costs, difference perceptions of responsiveness, and threats to security). 2) We use these findings to critically evaluate the “access, anytime and anywhere” construct shaping the design of future technologies. We present an alternative vision called deliberate interactions and discuss ways ICT can support this online usage behavior
23. Acknowledgements We are grateful to out participants for sharing their stories. Special thanks to our reviewers whose feedback helped strengthen our paper. A grant from the Intel Research Council supported this research. Thomas Smyth Marshini Chetty Paul Aoki (Intel) Beki Grinter
24.
Editor's Notes
BREATH, SMILE Hello I am Susan Wyche and I am finishing my PhD in GaTech’s Human Centered Computing program. This afternoon I will talk about Deliberate Interactions – an interaction style my colleagues and I developed based on fieldwork I did in Nairobi, Kenya a few summers ago.
I will begin by motivating this work… In infrastructure-rich settings- LIKE THE UNITED STATES- designers expect widespread Internet connectivity in users’ homes and workplaces. (BREATH) Even outside of these locations, there are growing expectations that users will have high-bandwidth connectivity on their mobile phones and laptops and be able access the Internet in coffee shops and on university campuses. (PAUSE) This idea of “access, anywhere and anytime” shapes usage models, advertising and how our community imagines and designs future INFORMATION AND COMMUNICATION TECHNOLOGY of ICT.
But we don’t know if this access, anywhere, anytime model is appropriate outside of INFRASTRUCUTRE RICH COUNTRIES and in the “developing world” or the bluer regions ones on this map. . . That is what we explore in our research.
Specifically, in Nairobi, Kenya. I chose to go to Nairobi because it epitomizes the fast-paced technology adoption evident in other major African cities. Other practical reasons such as English is widely spoken, and Kenya is relatively politically stable country–I should note that this research was conducted prior to the unrest that followed the disputed presidential election in Dec. 2007.
Also, Nairobi is home to an established professional class. Prior research in Human Computer Interaction for Development or HCI4D tends to focus on users living in rural areas, with low income and education levels. This research has been instrumental in providing knowledge about how to apply HCI principles to the design of technologies for developing regions. But we saw an opportunity to learn from individuals who regularly use ICTs in a “developing country.” So we interviewed and observed users who interactions with computers, the internet, and mobile phones are shaped by very different social, economical and technological contexts than what has been previously been presented at CHI. Specifically , our participants lived in an urban area, were literate, college educated, many had travelled outside of the Kenya, some had spent time in the US, AND they were higher income compared to most people in their country, and most important our participants had access to a broader range of ICT than individuals examined in prior studies and regularly used them in their daily lives. We chose to study professionals living and working in Nairobi who regularly use ICT, because understanding how they use technology highlights constraints people encounter when using the Internet in infrastructure-poor settings- like Kenya.
Just one more thing about what motivates this work and speaks to the timeliness of it. Prior work in HCI4D predominately focuses on the mobile phone, recent developments such as the arrival of an undersea fiber optic cable promises to usher in a “new Internet era” in Kenya and other sub-Saharan African countries, but there is very little research examining users’ experiences using the Internet on desktop there.
That brings me to the outline for the rest of this talk. PAUSE, SMILE, BREATH
To understand how ICTs were currently being used in Nairobi, I lived there for six weeks, and my collaborator Paul Aoki was there for two of them. Because of this short amount of time I employed rapid ethnographic field methods. This meant interviewing key informants, such as IT professionals. We conducted 2 focus groups with computer science students at area universities, and during all these sessions we asked participants to describe a typical day at work. I also asked questions about their access to technology, and their experiences using the Internet, email, SMS, television, and mobile phones. (BREATH) To complement these interviews and to understand how our participants used ICTs in their domestic lives, Paul and I conducted 10 in-home tours and observations at shopping malls, coffee shops and on university campuses. The findings I present here are based on 27 transcribed interviews, nearly 200 photographs and close to 100 pages of fieldnotes. The constant comparative guided our analysis, or myself and my collaborators-who are experts in this area- looked for key issues, and recurrent events in the data that became categories to focus on.
And we found four factors that shaped our participants use of ICT, : high costs, differing notions of responsiveness, threats to physical security, and limited bandwidth. (CLICK) BREATH and SMILE While other studies have discussed some of these findings, we were struck by how this access, anywhere, and anytime model didn’t hold true among our participants - Internet access remained tied to particular locations- it wasn’t everywhere - and participants ICT use was planned , it didn’t happen ANYTIME , we decided to call this “deliberate interactions.” Moreso than these findings, we really see this new interaction style and ideas for how to design for them as the contributions of our work.
We define deliberate interactions as a planned and purposeful interaction style that involves offline preparation. We really see this interaction style as one that is counter to how ICT are typically designed, counter to this idea of access, anytime, and anywhere. Rather than being unplanned and opportunistic as some researchers have described internet use in the “developed” world – in Nairobi it is the opposite and a great deal of planning and preparation is done before using the Internet and access is tied to individuals’ location.
For the sake of time, I am going to talk about two of these factors and connect them to our idea of DELIBERATE INTERACTIONS. Those are HIGH COSTS and different notions of responsiveness. I encourage you to read the paper to learn about Limited Bandwidth and Security.
I will talk about these findings and at the same time provide ideas about how to design for deliberate interactions. PAUSE, SMILE, BREATH
Participants consistently explained how the high costs of Internet access affected ICT use. More than half our participants had computers in their homes, but they just were not connected to the internet. High costs made having Internet access at home unaffordable for all but two of the nearly 30 people we interviewed. Those participants who did not have internet access at home didn’t think getting it justified the cost particularly when they had access at work. Most participants told us they arrived early or stayed late at work to use the internet for non-work related purposes. SEE THE QUOTE HERE (DON’T READ) (PAUSE)
Additionally, many of our participants told us that visiting an internet café, or what they called a “CYBER CAFÉ” was more economical than purchasing a residential Internet access plan. Most of our participants lived within walking distance of a café told us they visited them when going to work was not an option. BUT - High costs also affected how our participants used computers at cyber cafes. They rarely talked about surfing the internet or spending long periods of time online because they were paying for use by the hour, this was an issue among our participants, who were by all measures higher income than most of people in their country. Further, the reality is that speeds available in such cafés, where a single dial-up connection is sometimes shared between multiple users, can be painfully slow. Thus, participants are forced to experience the Internet at a glacial pace and pay for each minute they spent waiting for pages to download. So it was in our participants’ interest to carefully monitor how much time they spent using the Internet.
Consequently our participants made a point of using their time online as efficient as possible. A common strategy for maximizing this time was to prepare as much of the necessary documents for online interaction at home prior to going to the café. For example, participants told us they would compose an email message at home - on an offline computer- save it on a USB flash drive, and then visit an Internet café to email the message. This eliminated the need to pay for time spent composing an email. Other participants told us they quickly looked for websites at Internet cafés then saved them to a flash drive to view them later on their personal computers at home. This participant’s quote was representative of many we heard: “ I think the flash disk is one of the best things to come out, because before, I used to save my research on a floppy disk, the capacity was very small and by the time I saved one or two web pages it’s full, so my flash disk helps me to download much of the info I want, then I go home to analyze them .”
I want to present some ideas about how we can begin to design for this interaction style because while Kenya’s internal and external connectivity is improving steadily, the issues of high costs are likely to persist in the near future. First , One route to this goal is to maximize the usefulness of users’ limited time online . Systems researchers have proposed technologies for coping with slow connections using techniques such as pre-fetching, delay-tolerant store-and-forward networking and proxying, and the use of unconventional network transport methods such as USB flash drives and data “ferry” computers . These technologies operate at the systems level, seeking to preserve existing application software intact while modifying the plumbing underneath it. BUT, it may be more productive to modify the user interface. For example, one might give the user interactive control over how limited bandwidth is used. Current browsers download content indiscriminately , offering a few controls such as downloading all images versus downloading none. Second, A prominent phenomenon in our data is the offline computer — a computer that is capable of being connected to the Internet but is not connected due to cost constraints. (BREATH) We suggest that user interface could support better use of these offline computers by simplifying the shuttling of application data between online and offline computers. Interactive tools for creating offline Web browsing caches or offline email stores are a start in this direction. However, recent experience with online/offline versions of Web search and other collaborative Web applications indicates that they will often need to be split into a fully interactive front end and an asynchronous back end, thus requiring changes to both interaction design and underlying application semantics. (BREATH) This suggests the need for user interface toolkits to ease development of applications with this kind of split interaction. Finally, A third route is to design to exploit the different access and cost structures of the networks available . Because access to network bandwidth is limited and tied to specific locations, we suggest placing “media kiosks” in suitable locations where users can download content onto flash drives and use it on their computers at home.
Second finding I will discuss, is what we decided to call responsiveness . Designers of collaborative systems typically assume that communication occurs between individuals and groups with similar Internet speeds and connectivity. (PAUSE) Our participants who regularly received email from coworkers in countries with widely available and high-speed broadband connections told us their lack of constant Internet connectivity affected how quickly they could respond to messages these users sent. One employee of a multi-national NGO told this story:
“ This week, I was working at home . . . and I missed email. I had 174 messages, 174 messages! From South Africa, from Washington D.C., and I have to reply and send things. I don’t have Internet at home, I can not respond to emails as quickly as they can.” This was similar to what we heard from several participants. When they weren’t at work, they were unable to follow the ongoing email threads taking place among their coworkers because they did not have Internet access outside of the workplace . (Pause) Consequently, they felt overloaded. Internet access caused their email response times to be longer than their coworkers in places like the United States who had constant and easily available Internet access. Sometimes to make up for dealing with slow connections or because participants realized they would not be able to respond when not at work, participants made alternative arrangements to keep their responses timely by calling their co-workers, but this was also frustrating.
Managing different expectations for responsiveness and alerting those in more connected environments about their colleagues’ intermittent Internet access was a constant struggle for our participants. And they were also concerned that their colleagues overseas would view their delayed email responses as an indication of a poor work ethic rather than a reflection of their constrained ability to connect to the Internet. Again rather than being anytime and anywhere our participants’ mail use was very tied to a location — they had to be at work or at an Internet café to communicate with coworkers abroad. Again, this meant that when our participants did go online they engaged in a deliberate style of interaction.
Finally, some participants told us they were uninterested in adopting a work culture that relied so heavily on fast-paced email exchanges, rather than face-to-face communication. (PAUSE) Discussions about the “cultural bias” embedded in ICT have been presented elsewhere and reported when ICT were first appearing in Kenya. Our research suggests the need for longer-term fieldwork investigating how western professionals’ attitudes about efficiency and ICT use may conflict with Kenyans’ cultural values.
Again some ideas for how to design for this interaction style. Given that Kenyans will likely continue to have inconsistent access to the Internet, those working in transnational organizations will continue to feel frustrated by the mismatch in responsiveness between coworkers in “ developed” countries. Collaborative tools and email clients can be altered to help reconcile differences in response times to alert user with higher connectivity speed and easier access to the internet about co-workers with deliberate interaction styles. However, additional contextual factors beyond those considered in existing “responsiveness” tools may be required. (PAUSE) A simple idea along these lines is illustrated by the Google Mail feature that exposes the sender’s time zone.” (BREATH) More generally, email clients could maintain data on indicators such how promptly users respond to their email, the time distribution that a computer has network connectivity, or the quality of the Internet connection over which email has been recently downloaded. Such data could be shared with users in various ways, like visualizations of response times or schedule information.
To summarize, I present results from a qualitative study of how professionals living and working in Nairobi, Kenya regularly use ICT. There are two contributions of this work for the HCI community. First, we provide empirical evidence demonstrating constraints our participants encountered when using technology in an infrastructure-poor setting. These constraints are limited bandwidth, high costs, differing perceptions of responsiveness, and threats to physical and virtual security. Second, we use our findings to critically evaluate the “access, anytime and anywhere” construct shaping the design of future technologies. We present an alternative vision called deliberate interactions discuss ways ICT can support this online usage behavior. In the future I would love to build and evalaute some of our design ideas.