CAE Optimize Pairings & Rosters Improvements
Build, Measure, Learn Design Sprints
Process Overview
Discover & Define
‘Improvements Design Sprints’ follow a cut down ‘Double Diamond’ UX Model to speed things up without compromising too much.
First I facilitate ‘Prioritisation Workshops’, looking at previous research findings, pains of the current product or any business gains. Attendees might include a Product Manager, B.A., Lead Dev and UX Designers. Using a ‘MoSCOw’ framework (an acronym of ‘Must have’, ‘Should’ and ‘Out of Scope’) allowed us to remove waste and write a clear brief for UX tickets.
Develop & Deliver
Consisted of a rapid two week design sprint. First week User Flows, Use Cases and Rapid Designs. Usually I like to rapid prototype, working with Design System components, testing the connections between screens. If complex, I present a simple screen flow marked up with notes.
Second week might focus on validating our assumptions. Testing either internally or with Customers. Or just getting a B.A. to review my work and pass back feedback.
Finally wrapping up the sprint with Design, Stakeholder and Developer Feasibility reviews. At this point I’ve completed a Jira ticket with user stories, prototype, mini-video demos, and design documentation.
#01: New Iteration Flow
Test Task
Ran remote design tests with Peach and PAL Airlines to test the setup of a Pairings Optimization.
Findings
From test findings Crew Controllers found it hard to know how to start.
Needed training to understand the process.
Outcome
Created a simple flow to guide a Crew Controller through smaller steps, surfacing key tasks to start a new Iteration.
#02: Rule Comparisons
Test Task
Tested Crew Planners understanding of an ‘Iterations Scenarios’.
Tested to see if they understood its purpose or how it might differ from others.
Findings
Participants got stuck in the current product on how scenario rules were configured or differed.
Takes time to click into each scenario from the menu and then review each rule.
Outcome
Improved the knowledge of scenarios goals by being able to compare scenarios rules from the same iteration, and highlight differences.
Highlighting differences could reveal why results are better or incomplete.
#03: Pre-assign Roster Gantt
Test Task
Add duties reserved specific Crew like Training or V.I.P. flights, prior to planning the Roster.
Findings
Seeing the results in an expected format proved easy to understand for Participants
Outcome
Removed the hassle of doing it an external system.
Quickly make edits without re-running an Optimization, which could take hours.
#04: Manual Pairings
Test Task
Ran customer surveys to get feedback on results concepts.
Findings
Crew Planners wanted to manually create Pairings for special duties or complex multi-day pairings.
Outcome
Prototyped MVP UX flow to allow ability to create patterns of flights into Pairings.
Helped win new clients like Virgin Australia and Indigo in India.