Background

A brief summary about the company

Project Information

Alternative travel plan recommendation function is built to help customers quickly find suitable alternative plans after their flight changes. From data tracking, we know that the current conversion rate is constantly decreasing. So, we would like to iterate this function.

Previous Implementation

Problem and Goals

Process

The re-design process lasted around 6 weeks.

Before Start

Three main flight delay or cancellation scenarios were analyzed.

After discussing with product lead, I decided to re-design the current recommendation page instead of re-constructing recommendation logics and algorithms due to the following reasons:

• The specific situation varies greatly.
• Impossible to recommend plans by users' locations.
• Some key data (e.g. travel time between the airport and the railway station) are missing.

Research

Methods

Then I conducted research to investigate current problems. All following data and insights were obtained from three research methods.

Tracking Current Data

    23275 samples

Online Survey

1403 samples

Qualitative Research

6 In-depth interviews

Main Insights

01. Browsing current recommended plans is time-consuming

• Users get overwhelmed by the number of available plans after getting into this page (Interview)
• Current filtering is not noticeable and not efficient (Data Tracking, Interview)
  

02. Not enough flight information to facilitate decision making

• Users would like to know information such as flight cancellation rate and estimated arrival time after flight changes when think about alternative plans (Interview, Survey)

Ideate and Prototype

Based on key findings from research, I made the following list of design decisions for redesigning.

Insights 01

Browsing current recommended plans is time-consuming

• Users get overwhelmed by the number of available plans after getting into this page (Interview)
• Current filtering is not noticeable and not efficient (Data Tracking, Interview)
  

Users' Goal

Quickly and easily find alternative plans after flight changes

Design Decisions

• Re-design the filter in details and classify filtering factors into different importance levels
• Making default choices to reduce the number of available plans at the beginning
  

Prototypes 01

Details

01

Position and content of filter

Put the filter on top of the page to make it more noticeable. Besides, Will show the departure time, arrival time, and travelling method as an overview

02

Detailed Filter Page

According to the tracked data, traveling method, departure time, and arrival time were the most clicked factors. Hence, I put them at the top of the filtering page. Nonstop flights only is the most popular method of travel, so making it as the default choice.

Assumptions

01

Position of filter

Putting the filter on top of the page could encourage more users to use it, which will be tested later to determine whether it is a good position

02

Sorting

Current default sorting order (from early to late, from low price to high) combined with default choices could reduce the number of available plans and hence no further sorting is needed

Insights 02

Not enough flight information to facilitate decision making

• Users would like to know information such as flight cancellation rate and estimated arrival time after flight changes when think about alternative plans (Interview, Survey)

Users' Goal

Easily find a more ‘reliable’ plan after original flight changes

Design Decisions

• Show some key data (estimated cancel rate, estimated arrival time) along with basic flight information

Prototypes 02

Details

01

Showing original flight information

With the aim of building a close-loop ticket purchase platform, original flight information will also be shown in the page for users who bought the ticket in our platform to change/get refund easily

02

Showing more flight related data for reference

Recent flight cancellation rate was the most concerning data. Other basic information including planned departure and arrival time, flight number, aircraft, and price are also presented. For the flight in the next 24 hours, estimated arrival time, another key data target users would like to know, will be shown.

First-round Evaluation

The design was on the right track but had to think more about target users' true needs.

Overview

Methods

I have conducted 10 in-person usability testing sessions with random flight data with target users who are on business trips frequently.

Overall Results

All participants were able to complete all core tasks within 10 minutes.

Key Insights

Details

01

Original flight info - low priority to show here

• Presenting entrance for change/get refund was not necessary. Participants mentioned that their focus was to quickly find the alternative plan.
• Sorting is needed. Current default sorting could not satisfy many users' needs in searching.

02

Low granularity of detailed filtering

• Time interval in filtering departure and arrival time is too large. Most users want to select the exact time if they are in a hurry.
• According to observations, most users would like to filter travelling method first since they didn't know the default setting. Current filtering was still not convenient.

Second-round Research

Although the first-round design did not have big usability issues, there was still space for improvement.

Observations

One interesting observation was that many users tended to filter the travelling method first. And It would take them around 15 seconds to think about the five options listed under the travelling method.

Assumptions

• Most users are familiar with ticket-purchasing page which has two tabs: Non-stop flights only, Transit.
• Additionally, many users are not familiar with the term" Air-rail connections".
  

Hallway Testing

Persona

Second-round Ideate and Prototype

Combining both first-round and second-round research insights, here comes the second-round ideation and prototypes.

Insights 01

Users are not familiar with the new layout we have in 1st-round design (Usability Testing, Hallway Testing)

Features 01

Maintaining the same hierarchy as ticket-purchasing page since customers are more familiar with the layout

Prototypes 01

Details

Maintaining the same lay-out as ticket purchasing page. Users could filter the recommended plans by traveling method at first.

Insights 02

Designing for business travellers. Their goal is to find the plan arrives before certain time point and would consider Transit (First-round Research)

Features 02

Only designing filtering departure and arrival time function under Transit Tab

Test & Iterations

Tested how to better filter the Departure/Arrival time and Travelling method under Transit Tab

Prototypes 02

Details

Under Transit Tab, Traveling Time and Traveling Method are put at the top of the page for easy filtering.

Insights 03

Presenting original flight information is not helpful when finding alternative plans and current filtering is still not convenient without sorting (Usability Testing, Observation)

Features 03

Deleting original flight information and combining the filtering and sorting

Prototypes 03

Details

Under Transit Tab, sorting by price, departure time, and total travel time are listed for efficient sorting and selection.

Outcome

New version of finding alternative travel plans was officially released on Jul 20, 2021. The following data was extracted on Aug 10, 2021.

Next Steps

🚪 The entrance for the page was not noticeable.

The exposure rate of the function was low. Should consider making multiple entrances.

📑 More details could be added.

There were several feedbacks regarding on adding more details to make the recommendation “smarter”. For example, Fast transit plan could not be opted for “advanced” business travellers... These functionalities could be added after conducting more research.

My Takeaways