Samasource

 
 
 

About Samasource

Samasource provides data services to large enterprise clients. They are a social enterprise focused on closing the gap between global talent and economic opportunity.

Design Goal

The VP Product wants to increase service quality provided by their agents. She asked us to design a qualification process that could be integrated into their existing Sama Hub online project work flow.

My Role

I led a team of 7 designers to execute this project over 4 weeks. I was the primary client contact and facilitated weekly client meetings and design studios.


Quick Synopsis

Designing a Web Qualification Flow

Given most of the team's unfamiliarity with this type of business and the culture in Kenya, we spent a lot of time talking to the client about who is the user, what environment the agents work in, and what their goals are.

From there we conducted market research to analyze key traits of successful learning environments done online. We identified common design patterns and incorporated them into our qualification flow.

We prototyped an MVP version to test with 7 users in Kenya over Skype sessions, and collected a number of insights. These drove another iteration of low fidelity prototype.

Our final deliverable was a high fidelity prototype including screens for alternative and edge cases. The Samasource product team is now integrating this design with their other learning product.


Research

A motivating learning environment

The most unique aspect of working on this project was learning about the mental models of microworkers in Kenya. We used a persona to guide our initial prototype, and we validated and invalidated our assumptions during usability testing.

Before we started sketching, we investigated characteristics of other learning environments. We turned to Ed Tech for some key indicators - looking at companies such as Khan Academy, Coursera, and Edmodo. We found that most of these organizations used dashboards, practice questions, and progress bars

The general theme is to encourage a growth mindset. The agents may be working, but they can learn at the same time.


Approach

Design Requirements

We started with a clear understanding of what functionality is needed in this process.

  • Check if they have the skills to start a project through a test
  • Build skills that enable them to complete quality project work
  • Receive guidance on how to prepare for project work
  • Only work on projects once they have proven their skill level

Some constraints included:

  • Integration with their existing system
  • User unfamiliarity with technology
  • Connectivity in developing countries

Design

Task flows


Current Flow

Some pain points identified in the initial flow include:

  • Agents could access work for which they were not qualified
  • There was no critical path the agents were led down. Agents had free reign to access any projects.
  • Project managers had to spend a lot of time manually administering the Sama Hub workspace.
  • Production and Training projects were accessed in the same way, creating ambiguity around qualification.

Proposed Flow

The new flow focused on reducing complexity and adding a lot of guidance for agents:

  • New dashboard gives one clear call to action
  • Progressive disclosure of screens for agents based on different scenarios (i.e. passing qualification or not)
  • Agents take a test to unlock Production work; eliminating the need for Project Managers to manage permissions
  • New Project Home page clearly indicates which projects are Production vs. Training
 

Low Fidelity prototypes

Since no qualification process existed on Sama Hub yet, we gave ourselves the freedom to "big sky" ideas. We held several design studios to come up with multiple options for each screen, and we decided on the final screen designs through a process of divergence and convergence. 

In hindsight, I would have spent more time on user research to better inform our initial designs. For example:

  • shadowing an agent on how they currently use Sama Hub
  • leveraging conversation starters to understand what concepts and feelings they have towards their work and training
  • testing a cheap prototype by seeing how agents respond to tests (before building an entire process)
 
IMG_7417.JPG
 
 

Medium Fidelity Prototypes (Select)

From hand sketches, we transferred the converged designs onto Balsamiq. We used these to conduct the first round of usability testing with microworkers in Kenya.

 

Usability Testing

After conducting usability tests with 5 microworkers in Kenya, we found some surprising and interesting themes:

  • Users thought the "lock symbol" meant that they needed a supervisor to grant them permissions
  • Users liked the transparency of how far they had completed the test and production work
  • Users liked seeing the report card
  • Users wanted the training questions to show their correct/incorrect answers immediately after each question so they could learn as they went along
I don’t know what to do. I think I would call my supervisor.
If I don’t pass the test, does it mean I’m out of a job?
I like seeing what I got wrong and being able to learn from it.
 

High fidelity prototypes

Below is the flow an agent would experience if she did not yet qualify for the project, but does pass qualification.

Screen 1.png
 

Before & After Clickable Prototypes

Before

After


Retrospective

Going Global Takes Many Iterations

 
 

We completed two iterations of the prototypes, but learned a ton including:

  • Limit primary calls to action per page - For microworkers, time is precious and we want to enable them to quickly complete tasks without distraction.
  • Culture nuances are important for visual design - Symbols have different meanings like the lock symbol was thought to mean an administrator needed to give permissions. The original intent was for that to indicate users could unlock sessions through completing the qualification process.
  • Know the audience's mindset - Microworkers have a strong desire to build skills and they expressed wanting to see their weaknesses and be given an opportunity to improve. Therefore, the focus of designs was not on putting in gates and security measures, but rather giving opportunities to learn from mistakes.