From No. 3, Vol. XLVIII, “The Digital Dividend”, 2011

Information and communication technology is deeply woven into the fabric of society and is integral to the way we do business, entertain ourselves, talk to each other, learn about the world, and even feed ourselves. With nearly five billion mobile phones worldwide, the reach of ICTs is increasingly global. However, even with this near ubiquity, the benefits of ICTs remain uneven -- access to the "world information society" does not immediately grant membership.

There are many reasons for this. Resource-constrained countries are struggling to provide electricity and connectivity to remote villages and rapidly expanding urban areas; ICTs for schools and offices remain seriously underfunded by local governments working to provide basic services; and no matter how compelling the latest personal computers, tablets and smartphones may be, many families and businesses simply cannot afford them. Yet economics and affordability are not the whole story. There are other, less obvious factors, that keep many people from using and benefiting from ICT, simply because most ICT is designed by and for the wealthiest 30 per cent of the planet. Among these factors are: constraints in education and literacy; an incredible diversity of small language communities; political, religious, gender and other social prohibitions on use; and differences in cognitive models -- how individuals frame and organize information about the world. These issues cannot be addressed simply by throwing faster or cheaper technology at them. Good design in any context requires that solutions be tailored to the needs and desires of the people who will use them and this, in turn, depends on a deep understanding of the context and constraints in people's lives. A central problem with ICTs for the poor is that this kind of understanding is often neglected.

The examination of context and designing for constraint are core topics of research for the Technologies for Emerging Markets group at Microsoft Research India. For the past six years, we have been researching issues surrounding ICTs for global development (often abbreviated ICT4D). This article briefly highlights two of our projects, illustrating the importance of matching technology and constraint. In each case, the design of the technology is intrinsically connected to the context, needs, and capabilities of the end users and the aid organizations we are working with.

EXTENDING PEN AND PAPER TO THE DIGITAL WORLD IN MICROFINANCE

In India, as in many places in the developing world, paper remains a key component of information management. Visit any government agency, hospital, or school in India and you will see masses of paper and carbon copies used to keep track of almost everything. While the problems associated with paper are well known, paper brings with it many advantages. It's cheap, it's language-agnostic, people are very comfortable working with it, and there is a sense of permanence to paper that the intangibility of digital information cannot match. Our question: Is it possible to combine the appeal and simplicity of paper with the many benefits of digital data manipulation?

To explore this question, we decided to focus on banking and other financial services for poor communities. An estimated 3.5 billion people worldwide currently lack access to formal financial services. In India, self-help groups for microfinance (SHGs) have been very successful in providing access to such services for poor, unbanked people. More than six million SHGs bring formal savings and credit services to more than 86 million rural households in India. Women meet on a weekly or monthly basis in groups of ten to twenty and pool their savings in small amounts to borrow cheaply both from the group's own accumulated capital and from a linked bank.

However, one serious problem that SHGs face is the lack of reliable financial data management. In one common arrangement, "federations" of 150 to 200 SHGs work with an accountant in a nearby town to update their records on a weekly basis. Unfortunately, the weekly cycle of sending paper records to and from the accountant is often disrupted by bad weather, delays in error reconciliation or other issues, causing problems with getting timely information to and from the SHGs. An error in data entry by an SHG can take weeks to correct and cost members dearly in missed interest or penalty payments.

In a project led by researcher Aishwarya Ratan, we worked with a group of SHG federations in rural India to devise a solution that could be used effectively by low-literate, low-income clients with very little training. We created an application on an inexpensive handheld digital slate that accepts handwritten, pen-based ink input on ordinary paper forms and provides immediate visual and audio feedback in the local language. In addition, it simultaneously creates a digital version of the data that can be sent on to the accountant while leaving a paper version for the SHG. We tested this system in two field trials with 201 SHG members in West Bengal and Orissa and found that it substantially improved data accuracy, data completeness, and process efficiency.

We believe that two features were particularly relevant for the system's success. First was the production of the paper copy of the account information along with the digital. This physical copy of the records is important because it provides insurance that the information is always available. This sense of permanence is very reassuring for people who have little experience with digital systems and live in harsh environments where device failures are not uncommon. Second was an automated audio feedback on dues and balances for each member. It is important to note that many SHG members are non-literate. Indeed, SHGs often hire a literate person from the village to keep their records for them. The audio feedback was important because it provided each member immediate information on her current financial situation at the end of the meeting instead of having to wait for the round-trip to the accountant in the town to verify it. Just as important, this audio feedback helped to instill some trust in the system; there was no guessing whether or not the record-keeper had correctly entered their information. This second point highlights the importance of orality for many areas in the developing world. Most ICTs have been designed by and for literate societies. But what happens when the targeted users of a system cannot read? Another project from our group uses oral interaction and normal speech on mobile phones to bring information and a voice to people in rural India who have a difficult time communicating among themselves or being heard by others.

SOCIAL MEDIA FOR RURAL INDIA VIA MOBILE PHONE

When most people talk about a "world information society," they are implicitly talking about the Internet. However, universal access to the Internet remains far away for many countries. In 2010, Internet penetration in India was about 6.9 per cent. In contrast, mobile penetration in India stands at 68 per cent and is growing fast. For India and many other places in the global south, the mobile phone is the first and likely only ICT that people want or need, not counting television or radio. Cheap devices, cheap call rates, and an oral medium are keys to the success of the mobile. We have been very interested in exploring systems that let people use inexpensive mobile phones with their own language to share news and information with their local communities and the wider world.

One huge challenge for ICT design in India is the dramatic linguistic diversity; there are hundreds of languages and dialects spoken here. For example, in the tribal regions of Chhattisgarh, India, there are more than five million speakers of Gondi, Kurukh and Kui. However, there are no news outlets (print or audio/visual) in these tribal languages, and very few tribal journalists. In addition, it is illegal to broadcast news over community radio in India. We wanted to know if we could build a system where tribal citizens could share their own news using only their mobile phones. In partnership with CGNet, Massachusetts Institute of Technology and the International Center for Journalists, researcher Bill Thies built CGNet Swara, a voice-based citizen journalism system for tribal people in the state of Chhattisgarh.

Using this system, anyone can report news, issues, complaints, etc. in their local language by calling a number and leaving a voice recording. All submissions are reviewed over the Web by trained moderators to guard against spam, inappropriate content or noise, and are then published to the system. Once there, recordings are available for playback over the phone by local callers or for browsing online. Web access is not that relevant for local users, but it is important for making the content available for wider distribution and traditional media.

Over the past year, the system has received thousands of calls and published hundreds of reports. About half of all posts are concerned with governance and the redress of grievances: issues such as workers not getting paid under a government employment scheme, problems with rural schools or health centres, corruption, or teachers going without pay. Other posts relate to local news and events, including reports on pollution and public health emergencies (e.g., arsenic in water sources, outbreaks of cholera or diarrhea). And finally, a surprisingly large number of posts are tribal stories and songs.

As noted above, one important function of CGNet Swara is to make local news visible to traditional media outlets. An example of this is a tragic story that was posted to CGNet Swara about a father who was travelling around trying to get back wages from the government. While he was off trying to collect his wages, his son died in the hospital. From CGNet Swara, this story was picked up by the British Broadcasting Corporation, and soon thereafter by the large Indian daily newspapers, The Hindu and The Times of India.

A key facet of this system is that anyone who can use a mobile can use CGNet Swara. All interaction and navigation is done via pushing keys on the phones (input) with voice output and a very simple interface. For non-literate users in tribal areas, this is enormously important; it allows people who have hitherto been excluded from the "information society" to experience it in terms that they can understand and benefit from it.

The technology developed for both of these projects was completely determined by specific needs of people living in a context different from the researchers at Microsoft Research and, very likely, you. The keys to the success of both projects lay not in the technology per se, but in understanding how to fit the technology into the specific needs and capabilities of the people who are using it. Information technology has the potential to benefit huge numbers of poor people throughout the global south, but only if it truly matches the reality of their needs, and not just what we imagine those needs might be. Without this understanding, ICTs are little more than bits of metal and plastic.