Prepping for Launch in Kenya
Joe Wheeler and Peter DiCampo
What Went Wrong? is a participatory journalism project seeking to bring the voices of aid recipients to the forefront of the global discussion on foreign aid. We were fortunate enough to receive a data journalism grant from impactAFRICA in January 2017 and have been working to push our project forward since then.
One of the largest (and ongoing) challenges for our project has been choosing which technology to rely on for collecting citizen reports. This project began as a photojournalism project, but it quickly became apparent that traditional media would not suffice; we wanted to raise the voices of people who are vocal about foreign aid’s impact in their own communities, so we switched to a participatory journalism approach. While social media allows journalists to reach people internationally with unparalleled ease and affordability, it excludes potential users who might not have reliable or ongoing internet access. As many aid interventions affect communities that may not have internet access, the success of our project is dependant on our ability to connect and communicate with citizens from those communities. While mobile phones (and SMS in particular) have become increasingly common across Africa over the past few years, smartphones and reliable mobile internet haven’t reached the same level of penetration.
When designing our survey system for collecting citizen reports on international aid projects, we chose to use a mobile survey platform, Echo Mobile, that offers SMS and IVR (Interactive Voice Recordings) surveys to collect reports from citizens (at no cost to the user). Choosing mobile over social ensures accessibility regardless of internet access. The downside is that a mobile survey must be set up using a phone number (or numbers) within a given country, so it limits our reach to one country at a time. While our long term goal for the project is to start an international conversation, we chose to move forward with with this more accessible technology and launch our initial pilot over SMS, exclusively in Kenya.
With all that in place, we planned a launch date for late August, a few weeks after the 2017 presidential election, or so we thought. As most people know by now, the initial August 8th vote showed incumbent President Uhuru Kenyatta winning, but those results were contested by the opposition leader, Raila Odinga — leading to protest, riots, and eventually a national re-vote. Of course with such an active and pressing conflict affecting people’s daily life, this was hardly the time to start a project like ours.
Our team used that time to continue to explore other potential opportunities for the project, and we came across Typeform Chat (launched in March 2017). While we had discussed the potential of a chatbot from the onset of the project, the cost of building and maintaining a custom bot was prohibitive at the onset of the project. Typeform Chat, however, was affordable, and a Facebook Messenger bot would be accessible internationally at no extra cost; we were excited by the idea of integrating our survey system with such a popular messenger app, which boasts over a billion total users. Unfortunately, before we were able to build or test a chatbot, Typeform suspended the Chat feature. That said, we are exploring other comparable products (like Chatfuel, surveybot, or FlowXO) that would enable us to collect reports from citizens over the Facebook Messenger app. (And if anyone reading this happens to have experience with any of these platforms or similar ones, and has any feedback, we’d love for you to get in touch!)
The affordability and ease of use of some of these Chatbot template platforms has opened up an amazing opportunity to test and learn from a large scale implementation of both a mobile phone survey and an online / in-app version. A / B testing the two different options for our survey will begin to reveal the strengths and limitations of both versions, as well as help inform our decision around which platform is the right balance of accessibility and effectiveness for cultivating a large, multi-national, conversation. We are now adapting our timeline to a February launch with the addition of the Facebook Messenger chatbot, collecting reports via mobile in Kenya and via Facebook in other countries in the region.
By offering both an SMS version and a Facebook Messenger chatbot we know we will be able to reach concerned citizens regardless of their internet access while simultaneously establishing a future-facing, smartphone-friendly system. Ideally, as smartphone penetration increases throughout Kenya and the rest of the continent we can smoothly transition from our multi-modal system to a single, simple, and accessible online version — ensuring maximum accessibility for all-potential users in the meantime.
With these advancements in the project, and ongoing dialogues with community radio partners on-the-ground across Kenya, we are very excited to be launching our survey system soon! Many thanks to everyone at impactAFRICA and Code for Africa for their guidance and support. We can’t wait to start sharing out citizen reports and shining a light on the human ramifications of international aid in just a few weeks!
Follow us on Facebook and Twitter to stay up-to-date on the project.
