Jingru Guo

Designing Workera Test

Workera is a Deeplearning.ai company that focuses on bringing talents to the field of AI by providing mentorship and job opportunities that match their aspiration. Currently, Workera is targetting users who already have some technical knowledge, either professionally trained or self-taught, and are willing to get into the field of artificial intelligence.

As a product designer, I worked on launching the first version of Workera. When I was first brought on to the team, the engineering prototype was built. To use Workera, users first sign-up, complete tests, study feedback and improve, and eventually get connected to job opportunities that match their capabilities and aspirations.

Workera works like a funnel, as users need to go through the initial sign-up, test, feedback to unlock opportunities. Therefore, it is crucial for design to make sure we make the steps smooth and easy to minimize dropping off, and the test section is the first part of the product experience we need to work on.

Problems

When I first started, only 34% users started tests and 23% completed them, which was not ideal. The current experience requires users to complete the 7 tests, which were designed to give users a comprehensive evaluation, before they can move forward.

After digging into the current experiences and talking to users, I learnt severals things that attribute to the problems:

  1. Taking tests upfront is intimidating. Even for experienced professionals, taking 7 tests is a lot to ask, and lots of people don’t feel comfortable doing it, especially on topics they are less familiar with.
  2. Taking 7 tests is a big commitment. In total, taking tests could take more than 2 or 3 hours. The lengthy process and not knowing what they could gain after test completion made them less eager to finish the self-assessment.



Design Goal

The overall goal is to increase the test completion rate. To achieve this goal, we want to make the test process easier for them to start, as well as simplifying the test process to make it easier.



A less intimidating test

1. Minimizing mandatory tests

The previous Workera experience requires users to complete all 7 tests, which could take 2 or 3 hours, to move forward to feedback, and we want to simplify it.

Previous test flow



In general, Workera can help people find job opportunities in four types of roles: data scientist, machine learning engineer, software engineer with a focus on machine learning, and software engineers. As different roles require different skills, taking tests on a subject that is not required by one’s role is not very helpful.

To make testing more flexible, I added opt-out options for subjects that are not required for all job roles and moved the video interview to a later stage of the experience so that users don’t feel pressured when they first started. Through Inspectlet, I was able to see that the majority of users were using the opt-out option.



2.Creating practice test to get users started

One of the things makes Workera unique is the tests we are providing, but taking tests is especially stressful when you are not familiar with the test format and can’t prepare for it.

From Coursera’s product success, our team learnt that giving users something easy to get started could increase their confidence and motivation to continue their studies. Our data also showed that users who have started the test would be more likely to finish it, so I created a practice section with two simple questions. Users will be prompted to take the practice test after they have finished signing up.

Practice Test Final Design - Get Started



Practice Test Final Design - Completion

See progress as you go

Many users mentioned that there is a lack of test information, guidance and rewards in the previous design, and a huge part of the problem came from the interface design. In the final design, we added the progress bar, test length, test content and next steps onto the test dashboard. Users will be able to see how far they are to unlocking feedback and opportunities, as well as be able to opt-out of two sections if they are not related to users' skillsets.

Screenshot of the previous test experience


On the previous design, many users mentioned that there is a lack of test information, and, as they were going through the lengthy tests, it’s unclear what they would get in return. To provide users with more information and instant feedback to keep them going, I redesigned the test dashboard with a fresh new look, a progress bar, and more test information.

Dashboard Final Design - Test in Progress



Dashboard Final Design - Completed Test & Retake

Impact

The redesign successfully increased test starting rate by 97% and test completion rate by 70%.

Impact from Redesign