tiket.com / Bus & Shuttle
Improving Search Result Experience
This project will demonstrate my expertise in these three crucial domains:
Product Thinking
Product Design
Human-Centricity
In A Nutshell
Small Improvements, Large Impact
Due to multiple user Error Search Results, my team and I conducted the following to improve our design & user experience:
User Behaviour Analysis through Amplitude
Usability Testing on Bus & Shuttle Users
Impact x Implementation Matrix
How Might We Framework
These efforts resulted in a:
+4.2% increase in Search to Book CVR.
+7.4% increase in No Route to Choose Product Click-Through Rate (CTR).
+12.4% increase in No Route to Book Product Conversion Rate (CVR).
Here’s how the story goes ▼
Background & Context
Low Conversion Rate
tiket.com is a South-east Asia’s Travel unicorn, serving 52+ millions active users with various travel solutions like Flights, Hotels and Trains.
Following the release of our Bus & Shuttle product in early 2023, we identified a low conversion rate of 6.44%, prompting our team to investigate contributing factors and develop effective solutions.
Research
Multiple Problems, Minimal Effort
To address our low conversion rate, we conducted a comprehensive analysis using two research methods which lead us to finding multiple potential issues.
User Behavior Analysis via Amplitude's real-time data analytics
Live Usability Testing to identify user pain points.
By employing an Impact x Implementation Matrix (with Product Managers, the Business Team, and the Tech team), we prioritized the issues worth solving and concluded that the most critical and feasible issue identified was the frequent occurrence of "No result Error" on our Search Result Page.
The Root Problem
Inaccurate Location Data
Our Bus & Shuttle users often encounter a "No result Error" due to inaccurate location data from third-party aggregators.
The Cause
Vendors can define their own departure and arrival locations, causing multiple entries for the same place.
The Effect
This inconsistency results in users finding no search results when looking for locations from different vendors.
No Result Errors
Deep-diving into the Errors
Users will receive two different types of errors depending on what they search:
Error with Alternative Routes
Error without Alternative Routes
Design Evaluation
Issues & How Might We
🚧 Misleading Header
When users search for a specific terminal, the header only displays City to City.
→
❓ How Might We
Present accurate information to avoid misleading users.
🚧 Huge Error Message
The large error message overwhelms users and obscures alternative routes.
→
❓ How Might We
Reduce the error message's prominence to avoid misperception of search results.
🚧 Unclear Alternative Routes
There's no information displaying what the recommended routes are based off.
→
❓ How Might We
Display relevant alternative routes clearly and seamlessly.
Solutions
Small Improvement, Large Impacts
We improved 3 main aspects of our Bus & Shuttle Search Error Result Experience.
1
Header: Precise Search Results
Instead of a generalized "City to City" header, we now display the exact search terms entered by users to avoid confusion and misinformation.
2
Error Message: Reduced Error Perception
We redesigned the error message to occupy less space and appear less like an error state, reassuring users that their search results are not incorrect.
3
Alternative Routes: Seamless Recommendations
By clearly indicating the recommended routes (e.g., City to City), users can easily understand which Bus & Shuttle terminals are being suggested.
4
Before, After & Imapct
To validate whether my designs worked, we conducted an A/B Experiment to test both designs. The new designs resulted in the following results:
+4.2%
Increased CVR from Search to book
+7%
Increased CTR from No Route Error to Product Detail
+12%
Increased CVR from No Route Error to Purchase by 12%
Reflection
What I've Learnt & Special Thanks
During this project, I've got the opportunity to:
Learnt more about User Behaviour Analysis (Amplitude)
Leading Workshops & Using "Impact-to-Implementation" Matrix
Working with Business Intelligence Team
Conducted A/B Experiment and having the new designs win
However, although this project is deemed successful, the best solution would be to improve our location datas.
Shoutout to all collaborators
Huge shoutout to everyone I collaborated with during this project
Product Design team - Ule, Nindya, Yossie, Ryan S
Product team - Agung, Alwan, Hanifa
Business Intelligence team
Tech team
and many more others.