SAN FRANCISCO — Last week, at an event on artificial intelligence for economic development, Josh Blumenstock, an assistant professor from the University of California, Berkeley, gave a talk entitled “Fighting Poverty with Data.”
Blumenstock talked about his work gathering mobile phone data from the cell phones of users in Rwanda and then turning that information, such as the number of calls made per day, SMS volume, and international contacts, into interpretable metrics. His research is at the intersection of machine learning and international development, the focus of the event organized by the Center for Effective Global Action and the World Bank at Google’s headquarters in San Francisco. He highlighted how mobile phone metadata can predict poverty, and the implications of that in terms of targeting aid, crisis response, and impact evaluation.
At one point, one of the engineers in the room asked Blumenstock how location data factored into his work. “The problem is that people aren’t using smartphones yet,” he replied, matter-of-factly. While it might come as a surprise to people who have not spent time in developing country contexts, the majority of the people Blumenstock works with have basic phones they use only for calling and texting.
While smartphones are becoming increasingly affordable and widely available, their rates of penetration are still low in developing countries, particularly among the poorer segments of the population. There are 1.3 billion feature phone users, most of them in Asia and Africa, using devices that can sometimes play music, run simple apps, or access the internet, but lack the more advanced functionality of a smartphone. In low- and middle-income countries, there is a gender disparity in mobile ownership, and a number of barriers that stand in the way of mobile ownership and use, particularly for the very poor. These feature phone users have not upgraded to smartphones, switched from to 2G to 4G networks, or started to use data services in addition to voice and SMS, due to issues like price or literacy or relevance.
“It will be several years before rates of smartphone penetration are sufficient to imagine that the average person will be able to access smartphone-only apps.”— Josh Blumenstock, assistant professor at the University of California, Berkeley
“It's exciting to see all of the new smartphone-based applications being designed for low-income subscribers — many of which also seek to improve the quality of life of the user,” Blumenstock told Devex via email. “However, it will be several years before rates of smartphone penetration are sufficient to imagine that the average person will be able to access smartphone-only apps.”
Unless developers and designers focus on the short-term reality as well as the long- term future, they will fail to reach the relatively poor and marginalized segments of the population, leaving those who do not have smartphones even further behind.
Is there an app for that?
At the Mobile World Congress in Barcelona, Spain, last week, the software company KaiOS Technologies unveiled its plans to bring smartphone apps to affordable devices in order to better serve a market that other manufacturers have ignored. The announcement resulted from a partnership with Silicon Valley technology companies that will give feature phone users access to Facebook, Google Maps, Twitter, and more. Of course, it is in the interest of these social media giants to reach more users on the devices they use, but the majority of technology companies working on mobile innovation have shifted their focus away from feature phones and onto the next big thing, which limits the role they might play in developing countries.
“In the tech world, it is such an article of faith that you're designing for tomorrow’s problems, so to get anyone excited to design for what feels like yesterday’s problems is difficult and certainly not very sexy at all,” said Craig McIntosh, a professor of economics at the University of California, San Diego.
McIntosh said he is currently facing this challenge on a project in Uganda, funded by the United States Agency for International Development, that uses a digital food trading platform to link farmers up with buyers. The team is in the midst of a transition from research to scale up, and while the software firm that built the back end wants to transition to smartphones, the rest of the team argues it will be five years before these farmers have them. Even when they have the devices, their use of data might be limited, given the lack of access to free WiFi, high cost of data usage, challenges with charging, and more.
See more related topics:
Smartphone apps can succeed in developing countries, as Uber is demonstrating in a growing number of African countries where the service available, but they cater largely to the urban elite, McIntosh explained.
“Silicon Valley loves to present itself as solving the problems of the world’s poor, but is in fact focused on making money,” he said.
Based in San Diego, California, Qualcomm gets most of its revenue from chipmaking. While it is behind the entry-level processors that make some basic app access available on feature phones, it is more focused on building the foundation for fifth generation wireless systems, or 5G, which will bring ultra fast wireless speeds. Qualcomm says its inventions sparked the smartphone revolution, so it makes sense that its social impact program highlighted the role these devices can play in achieving the United Nations Sustainable Development Goals in its recent film.
“We don’t really design for dumb phones, because our whole focus is on leveraging mobile broadband access, and the opportunities created by that access,” said Michelle Martin, who leads global communications and marketing for Wireless Reach, a Qualcomm initiative that brings advanced wireless technologies to underserved communities. “A big part of the education and outreach we do does focus on the importance of people gaining access to smartphones. It is about driving those costs down. It is seen as a luxury good in a lot of countries.”
Other organizations testing the use of next generation mobile tools to achieve the SDGs include the U.N.’s World Food Programme. The team behind mVAM, WFP’s mobile Vulnerability Analysis and Mapping project, built a chatbot — a computer service that leverages artificial intelligence in a chat application — together with partners including Cisco, Nielsen, and Facebook. The idea was to respond to the growing popularity of smartphones and chat apps, and explore the potential applications of chatbots to their work, but they acknowledged when they launched the pilot that the majority of people using messaging apps are young, urban, and male, arguing that this would change as smartphone use becomes more prevalent.
Leveraging existing tools
Silvia Figueira, associate professor of computer engineering and director of the Frugal Innovation Hub at Santa Clara University, said that while she prefers not to use the dumb phone term, she agrees with the premise that the global development community should not get ahead of itself when it comes to the exciting potential of smartphones. At a recent event hosted in partnership with the African Diaspora Network, she talked about the work students at the Mobile Lab are doing to develop mobile applications for feature phones, or “not-as-smart phones,” as she often puts it. Figueira talked about her work with underserved communities and social enterprises to leverage existing tools and appropriate technology. In a follow-up interview with Devex, she explained it is not just about designing for basic devices but also designing for the constraints that users in developing countries face.
“I’d like to see companies spending more time at the last mile, more time in the field, actually designing for what people have at their fingertips now.— Sonali Mehta-Rao, co-founder and chief growth officer of Awaaz.de
“For mobile innovation, I think it’s really important for the design to be more user centric,” said Sonali Mehta-Rao, co-founder and chief growth officer of Awaaz.de, a communications platform that helps development organizations reach base-of-the-pyramid populations, and uses voice surveys, as compared to SMS, as a way to overcome barriers of literacy and language. “I’d like to see companies spending more time at the last mile, more time in the field, actually designing for what people have at their fingertips now. For us that was designing for dumb phones, because that’s what people have in their hands now.”
“When we talk about digital divide, the conversation often ends at infrastructure,” said Amina Fazlullah, a tech policy fellow at Mozilla who has worked in the United States and globally to promote policies that support broadband connectivity. But “once you get the network in the ground, and make the service available, there are still other barriers.”
In addition to addressing issues of cost, and helping communities that have long been offline to understand the value the internet can offer, global development professionals together with software development professionals can go beyond providing devices and developing apps with troubleshooting and other forms of support, Fazlullah told Devex. While it may be no surprise to anyone who works in a developing country context, it is less obvious to developers based in San Francisco or Austin that smartphones are not yet widely available. Anyone aiming to serve these populations should consider challenges such as consumer protection, as well as limited internet access or limited texting and minutes, Fazlullah said.
“I hope to get developers to think about digital inclusion at the development and design stage,” she said.
From spam messages to smart design
A recent paper by McIntosh of UCSD and a team of researchers explains how platform development that is driven by smartphone logic does not work for basic mobile phones. It is one of a number of examples of actionable guidance for anyone deploying ICT platforms for international development. Another is the USAID-endorsed Principles for Digital Development, which includes designing with the user. That means developing context-appropriate tools, working in an incremental and iterative manner, and ensuring the design is sensitive to the needs of the underserved.
"The reality is that you can have a reliable network connection, a phone in hand, the ability to afford mobile services, the knowledge of how to use a device, and the permission to do so, but none of that matters if the content does not resonate with, or the conduit is not applicable to, an end user,” said Chris Burns, director of the Center for Digital Development, within USAID’s Global Development Lab.
To have a dumb phone in a developing country is to be inundated with spam messages, McIntosh said, referencing the many SMS services that target people in these markets for everything from surveys to mobile health interventions. “Bad development engineering assumes one SMS coming from an unknown source will shift behavior in some way,” he said. In order to stand out from this chaos of unsolicited messages, organizations need to establish trusting relationships with users, who know to expect messages from them, which requires a fundamental shift in how organizations think about interface design, McIntosh added.
He said global development organizations interacting with users via dumb phones should consider using USSD, which stands for Unstructured Supplementary Service Data, rather than SMS. This technology works on a session-based connection, which addresses some of the challenges that arise when users go in and out of network. It also allows for decision tree logic, where users are presented with new choices based on what they input.
But while mobile money services have proven the most effective in leveraging USSD for communication between customers and their mobile payments platforms, he said, other examples of use are having issues with scale. It seems like every pitch competition focused on solutions for the poor features some kind of app to solve their problems. But when people first get access to mobile phones, McIntosh said, they tend to make calls to learn what they need to know, so the phone itself has already solved many of the informational needs they have.
“One of the difficulties of this whole world of information interventions is it’s easy to convince yourself people want to know something and to be wrong,” McIntosh said. “The key there is how successful have you been in identifying information that is usable, influential, and unknown to them, and it has to be that combination of things to have an impact.”
Update, March 12: This story was updated to clarify that the Principles for Digital Development are endorsed by USAID