Lessons from the Afriscout Usability Testing in Marsabit

By Dennis Muthuri
iHub Consulting
  Published 28 May 2018
Share this Article

You know a usability test is marked for greatness when the location of interest is 750km away from the city and the hired tour van breaks down on a major highway at 6:00am in the morning, forcing your team to push it against oncoming traffic to the roadside in the middle of a light drizzle. At this point, you realize that you've been delayed by 4 hours and require a miracle from your ancestors to cover 750km in under 7 hours.

That’s the prequel to the Afriscout usability study in Marsabit, something I tweeted about here, hereand here

First, Afriscout in 15seconds

Afriscout Usability test in Marsabit Initial sprint Afriscout initial sprint

Afriscout is an app that provides semi-nomadic pastoralists with satellite powered maps that shows the density of vegetation in their traditional grazing lands. The pilot project tested the potential of the innovation by delivering paper-based vegetation maps to pastoralists in Ethiopia and found a 78% adoption rate and 47% reduction in herd mortality in the first year. Paper maps revealed the effectiveness of the idea but to scale, the PCI Global team needed a way to bypass the handoff and give pastoralists immediate and direct access to the maps as they update and pasture conditions change. This is what birthed the idea for a mobile app. In partnership with Google, iHub Software Consulting and Hoefsloot Spatial Solutions for satellite images, the team conducted a one-week design sprint which resulted in a mock application that was used to validate assumptions with other Maasai pastoralists in Kenya. The mock application was received well by the pastoralists. The iHub Software Consulting team used the findings from the design sprint to build the application.

At the moment, the application is still being iterated on. We recently traveled to Marsabit to conduct a usability test with a select group of app users. The test sought to understand how users are interacting with the application in their environment and identify opportunities for improvement.

We learned a number of things when we were out in the field.

1. Continual exposure to different Mobile Apps sets the tone for familiarisation with your own app

Prior exposure to different apps sets the stage for success when using a brand new app especially when the patterns implemented are similar. It is only when you misunderstood what the hamburger menu did in a certain app, can you be proficient at guessing what it rightfully does when it’s spotted in another app. For this reason, designers are encouraged to use familiar patterns and only break them when certain we’re designing for the elite 1% of tech users or introducing something totally new. 

We noticed that users who had familiarity with prior apps such as WhatsApp and Facebook had already been educated on different UI elements. This is because the social value of the app outweighed the initial struggle to understand how it works. Afriscout’s iconography was well understood apart from cases of poor or non-existent user education by the community marketer.

2. Test Preparation is a necessity

"If you fail to plan, you plan to fail" is an age-old adage which is yet to be proven false. Gather the elements needed for the test in good time and share the plan with the team. Here are a couple of quick tips which helped us make the most out of the test.

  • Context: Test where your user is; Conduct your usability test in the day to day conditions your solution is used. This will increase your empathy for your users.
  • Use scenarios: Get the users to accomplish goals on the app; Come up with scenarios involving short goals to test the users understanding of your solution. This will help you identify opportunities to reduce the cognitive load for your users.
  • Brief the team advance: Have a test plan ready a day before and walk through it with the team members facilitating the interviews the previous day.

3. Develop for the users' connectivity

We have enough marketing sentiments that 3g internet is well spread and taking over the country, but the reality couldn’t be further from the truth. Developers and designers still need to think in low connectivity and have an offline first mentalitywhen it comes to creating consistent and predictable experiences for mobile and web.

During the entire test, 3G connectivity was only assured of in Marsabit town. However, 2G became more common once we got to the temporarily lush grazing lands of Moyale where our users were. During the test at Odda Primary School which is 250km from Marsabit, facilitators had to repeatedly walk to the highway to acquire connectivity in order to download the maps. Upon inquiry as to whether this was the norm, we uncovered that a regular 12-30km walk to the Moyale airstrip or a drive/hitchhike to Marsabit town was common in instances where better connectivity was desired. Realising that we hadn’t done a great job at communicating well in times of low connectivity, the iHub Software Consulting development and design team immediately made 2G development a priority for the app as their empathy for the users grew.

4. Plan for multiple distribution channels for your app.

Afriscout community marketer Afriscout community marketer

“Yes we understand that there are challenges in connectivity when using the app, but we found a way to use it, in life, you don’t stop because of challenges.”

Those wise words from our community marketer proved true when we realized that the major distribution channel of our app occurred offline via the Xender app, instead of online via the Google Play Store. Whether app marketers like it or not, Google Play works in as far as the internet is great but it’s unsuitable when low connectivity is a norm.

We found out that users can take up to 2 weeks to download the app via the Play Store due to a flaky internet connection. This would decrease the rate of conversions for the community marketers who after spending precious time educating them about the benefits of Afriscout, wouldn’t have the desired users download it as fast. This meant a slow start to experience the intended quick gains which would cause them to re-use the app. They opted to use Xender to send the app to the community members being onboarded as it was faster. This made the PCI team consider a different plan for tracking offline installations and activations.

5. Design for repeat use

Fewer steps, Greater wins. Two steps should be ideal a for a user to complete a goal on a mobile app. Any screen that needs to explain the purpose of the next screen should be re-thought if not removed first.

A slight agitation on the users' faces would be noted when they missed a step and had to repeat everything again. This happened mostly when entering payment details, forgetting them and having to retrieve them again.

Test Wins

Afriscout field test in progress Afriscout field test in progress

The PCI team found out that the pastoralists would be willing to pay for the app on an annual basis given that it provides real value and saves them scouting time when grazing season arrives. The iHub Software Consulting team, on the other hand, received a lot of app and flow changes for improvement. After all, what’s a usability study without some homework for the dev team ;)


iHub Software Consulting is a design-led company that offers best-in-class software development and design consultancy. Our services include user experience research, design thinking facilitation, and software development. Interested in learning more about our work? Talk to us on [email protected]

Further Reading

  1. Detailed write up of the first Afriscout sprint
  2. What should happen next after running a Design Sprint
comments powered by Disqus