A/B Test Design - Waze

I created three prototypes of the Waze app which, at the time, presented possibilities for helping Wazers discover information on vaccination locations and/or availability.

Me Kevin Towner•2 minute read

Breakdown

Timeframe:

  • • 8.11.21 - 8.13.21
  • • One week assignment

My Role:

  • • UI designer
  • • UX researcher

Tools:

  • • Figma
  • • Useberry

Inspiration/Resources:

Overview

The Problem

During the pandemic, when vaccination locations were unclear, many people were confused as to when, where, and how they could receive vaccinations to address COVID-19. Using Waze as a foundation for the prototypes, How could I design an A/B test that addressess COVID-19 vaccination confusion?

Specifications

The test consists of three separate prototypes that represent two different solutions, and the control variable(s). Testing for significant changes in completion time between tests and control.

  • • Test C - Control
  • • Test A - Bottom Navigation Button
  • • Test B - Popup

Test C

Control Screens

Test C has two screens which remain consistent across all three prototypes, and thirteen additional screens that represent the user searching for vaccination information.

Control screens Select image for Test C prototype

Test A

Test A screens Select image for Test A prototype

Small popup window

For Test A, I chose to incorporate a pressable popup notifying users of nearby vaccination locations. The two locations mentioned in the popup will be displayed to participants in the form of a list after being pressed. After selecting one of the options, participants will view geographic and parking information on the vaccination location.

Test B

Bottom navigation button

Test B is similar to Test A, except instead of a small popup window, participants will encounter a button on the bottom navigation menu. The button will bring participants to the two itemed list mentioned in Test A where they can select an option.

Test B screens Select image for Test B prototype

A/B Test Quick View

Control

C completion Test C results - Useberry

Test B

Test B screens Test B results - Useberry

Control

C completion 2 users quit, and 1 dropped off before completion

Test B

Test C screens 3 users skipped before completing

Results

No Significance

There was no significance in completion time or completion rate between the Control and Test B. In both prototypes, there were participants who did not complete the prototype by opting to skip completion. In one case, a participant dropped off.

What about Test A?

I failed to collect data on Test A due to a combination of technical issues and Useberry versioning issues. Other possible contraints included:

  • • Time limit
  • • Solo project

UI Components

Small components

small components small components

Medium components

medium components medium components

Large components

large components large components

Takeaways

Final Screens

all screens all screens

Going forward

It's important to fully understand the tools and their limitations before delving deep into a project. It would be interesting to conduct the A/B test with all prototypes involved and on a carefully selected participant pool. Some constraints included a fairly limited amount of time to conduct quality research before designing the tests, Useberry's versioning specifications, and limited time to gather ideal research participants.