Redesign Case Study:  
Flight Booking
on myCWT Mobile 
 

Background

myCWT Mobile is the world-leading business travel app, which includes multiple services for business travelers such as trip booking and management. This case study showcases the process we went through to improve the app's flight booking capabilities, from collecting initial feedback, Identifying top issues and solutionizing. 

1. The starting point 

When I joined CWT, their mobile app was already up and running . "myCWT" included multiple services and capabilities for business travelers, including the ability to book flights. 

CWT's travelers had some choices for booking their flights - online on a dedicated website or with our travel agents, It was a strategic goal for us to shift these users into our app, but despite a long release period, the app's flight capability wasn’t adopted by our users.

 Initial design  

Flight results screen

Flight filters

Search screen

  Initial Stats  

booking

penetration

0.7%

conversion

rate

12%

Monthly transactions

185

average

booking time

4.4m

NPS

(satisfaction) 

-20

2. Research 
 

As the UX lead for the project, I wanted to investigate the cause of the low adoption and satisfaction. Assuming there was a connection between the metrics, I explored traveler behavior when booking flights, and compared it to user feedback to identify the gaps between user expectations and experience. 

  Usability study  

We start by collecting user feedback to make sure the existing solution meets user needs. In this case we chose a usability study as a way to identify usability obstacles that could harm conversion, and any gaps in the experience. 

The method

  • In-person interviews with business travelers, at their office locations

  • Each participant was directed to book a flight for their upcoming trip, using our app

  • The participants described their thoughts and considerations as they carried out the tasks

  • Both the session and the phone’s screen were recorded on video.

  • 15 participants in London, Minneapolis, St. Louis & Chicago

"I need to be there by a certain time, and that's the most important thing for me"

“That's confusing, I'm not accustomed to seeing flights presented like that”

“I don’t like these bundles, I’m getting frustrated, and I don’t want to do it anymore"

  Survey  

Once I had some qualitative insights into users' main friction points in the app, I set out to quantify them with a survey, which I used to understand the business traveler flight booking decision process, concerns and needs. 

The method

  • Online survey created on SurveyMonkey

  • Screener questions were set in the survey to ensure participants met the target profiles

  • Participants were asked about their process and criteria when booking a flight for business 

  • 500 participants from US, UK & Canada

Main insights

  • Flight search process didn’t match users’ needs for time-based search

  • Flight grouping didn’t match users’ needs for comparison

  • Flight bundles limited user choices

Flight properties affect on decision

Flight time criteria

  3. Defining the goals  

Taking everything that we learned, we Identified required changes to improve the flight booking flow:

Ungroup Flights

present all flights on a list, to allow time-based selection 

Un-bundle Flights

Allow selection of outbound and return flights separately

​Add time filters

provide users with time exploration based on needs

 Main KPI's 

  • 30% conversion rate 

  • Reduce booking time to 2m

  • Improve user satisfaction (+10 NPS) 

  • 5K monthly bookings 

  • 3% booking penetration

  MVE  

We ran a short workshop with Product and Dev, during which we defined the product's MVE (Minimum Viable Experience) as a team. This was a way to define an MVP with the user experience in mind, meeting our users’ basic expectations, with a minimum level of experience. 

  • Defining the minimum level of experience we’d like to maintain/achieve in the first phases.

  • Shifting focus from features to experience.

  • MVE standards evolve based on user’s expectations (as opposed to MVP’s static definition)

4. Design iteration 

We started working on multiple low-fidelity concepts and prototyped them. Our goal was to run a high-level assessment of our design ideas before development. We tested our concepts with potential users of this app who matched our business traveler personas. 

After each test, we adjusted the design according to the feedback, and tested again. This iterative process allowed us to rapidly and effectively test our assumptions and improve. By the end of the design iteration process, we had a design concept that we were much more confident matched the users' needs. 

Methods and Tools

Un-moderated, online tests with business travelers 

  • Prototypes were built on Framer and Principle

  • Platform: Usertesting.com

  • Screener questions were set in the survey to ensure the target profiles were met

  • 100 different tests were conducted on 12 different concepts

Questions for time filters 

Before vs. after search?

Double vs. single scale

5. Wireframing & Definitions 
  Final  
    Design!   
6. User feedback loop 

As we launched the updated version of myCWT flight booking, we wanted to be able to constantly listen to our users and look for ways to improve for them. To do that we constantly collected feedback and acted upon it to improve the product and the user experience.

We used in-product surveys to collect contextual feedback from specific users, by asking the right questions at the right time via targeted dialogs with flight users.​

7. Impact

The myCWT flight booking redesign was well received and had a positive and immediate impact on all the targeted KPIs. It was released to a small batch of users, and once we were sure it was the right way to go, we released it to all of our users. We kept collecting usage data and user feedback, and improved the product as it evolved to become the world-leading business travel app.

Usage data: 

Conversion 

rate

28.5%

average

booking time

2.2m

NPS

(satisfaction) 

+9

Monthly transactions

8k

©2020 by Oz Richter. All Rights Reserved.