Bus Onda

Onda(온다) means oncoming and is also a local public bus notification app in Korea. Imagine when you go to work, study in the library, or it's freezing cold, an app enables those who are worried about missing the bus to bring a relaxing experience by adding a bus alarm to save your time waiting for the bus. May your life become more productive without wasting.

Product Design
& Design Strategy

14 mins Read



Onda had been initially developed for my friend himself. He learned the basic programming language from zero. And I was in charge of product design. A month after, Onda version 1.0 was born. The app had been improved slowly during the first two months of launch.

One day in late July, the number of people downloading Onda had increased tenfold. Thus, our server suddenly crashed for half a day while it was still being optimized. That day they could not use Onda and as a chain effect, gave negative reviews and low ratings.


The First Version of Onda
on May, 2019

We later found out that those people came from the App Store's Today recommendation. This meant we failed their expectations and missed the opportunity to promote Onda.

In this way, we decided to put in 100% effort to be ready to strike when the next opportunity presented itself. We had to take it seriously as a business project. Thus, the journey of Onda iteration began. 


Featured on App Store Today's Section
"Best Apps of July"

Business Goal

We roughly developed two key results, improving user experience to build stickiness into the product and growing revenue to afford pricey app development costs. My role was to compare multiple pricing model experiments and simultaneously lead the design iteration. 


At that time, no one knew how far Onda could go, and none of us had any experience in building digital products from scratch. With no clear product goals, no professional team workflow, and no existing experience in product management and operations, it was difficult for us to move forward.


But still, there is always a way out. The Internet provides unbounded possibilities for learning. Apart from that, just making multiple trials in practice. For Onda version 2.0, we kept running four existing pricing models test for a half year in support of our design results and metrics. We also collected user feedback and interviewed potential users to uncover deeper insights.


Pricing Model

Over the six months, We had used four different payment models for Onda. In the meantime, we learned to look at application operational metrics that we can use to better improve the experience. Each model produced different results in terms of revenue per user and so on. In the following, I will break them down separately.

With one accord, the developer and I chose the payment model which could balance both business value and user experience. We aimed people were happy with what they have purchased. Compared with a paid app, freemium provides optional choices and brings more access. Despite the higher revenue from the paid model, Onda with freemium saw a substantial trend in downloads and purchases during the fourth test. 

As more and more active users, Onda 2.0 was ultimately transformed into a freemium app, the subscription model plus offered free additional features. Yet to make people comfortable spending in our app was a tougher road than we imagined.


Free Model w/ Display Advertising:
This pricing model lasted two months (47 days to be exact). Onda is a niche navigation tool. The less viewed ads did not bring in enough revenue to cover the cost of maintaining the software. Also, adding ads that distract attention created a worse user experience. It r
educed user engagement and retention, especially with low-quality advertising.


$11/4pp(IAPs*) + $5(Ads)

Rev. per user

Paying users

*IAPs: in-app purchases


Paid Model:
Soon we removed ads and conducted the second test from 10 October until 7 December (58 days), it was a fully paid version with no ads. Although our paid app got some earnings into the pockets, monthly active users numbers were shrinking. A paid app did not give access to most people, who neither see value in the services that we provided. 



Rev. per user


Subscription w/ 14-Day Free Trial:
We tried to establish trust with people and let them try before they buy. In order to lower the barrier to download, we first considered providing a 14-day free trial from 8 December to 12 December (4 days). As a result of the third test, few people were willing to pay for the entire service when it stopped offering free trials, even if we offered the same quality as when they paid for it. 
Low retention rate shortly made us conduct another test again.



Rev. per user


Freemium w/ Subscriptions:
Since our design iteration made some progress, we separated Bus Arrival Schedule and Boarding Alarm into two main features. We decided to try the fourth revenue model and made Onda freemium from 13 December till the time of writing (19 days).

Bus arrival schedule was for free, whereas boarding alarm came to be premium. This meant to engage both paying and non-paying users. We assumed the earnings would be higher than the paid model if we kept and increased subscribers.


$8/8pp (IAPs)

Rev. per user

Paying users

Design Iteration Problem Statement

From field research, some errors of Onda notification were inevitable due to a problem with the national public data portal↗︎.

Product Goal

During the pricing model upgrade, more and more user feedback drove us to make a better bus notification application in Korea. Simplifying the bus panel and improving the usage rate of bus boarding notifications overtook the design direction of the revision according to user feedback and research.

The design revision focused on the following aspects — optimizing page information hierarchy, rebuilding the app’s identity, and improving the experience of bus boarding alarm.


Synthesis Research Results


The Research Team and DFunk Manager

The Research Team and DFunk Manager

Design Iteration

As constant learning by doing, we optimized rapidly page information hierarchy, after revising the journey map throughout insight into how people interacted with Onda.



lychee©How Might We

How Might We


Design Solution

lychee©User Approach





Unlike previous projects I'd done, I noticed it was not about focusing on the design details of a certain point and create inventive solutions. Inclusive design required me to look at the bigger picture. The team had to understand user diversity and determine what functionality and features The Welcome Box should include.

From initial findings to usability testing, research has played a vital role. In desk research as a researcher, rather than just understand Danish refugee context, it is important to go deeper into the full picture between the locals and refugees.

In addition, to make refugees feel truly included, the team needs to create an enduring, delightful, engaging product experience by understanding how refugees emotionally connect to Welcome Box kits.

In the next phase, our team will conduct numerous rounds of user testing to check experience issues and iterate through the cycle until the best experience solution is explored and then put back into development.



Seoul Downtown

Lychee Li

Hongbeom Park

Random pedestrians
near bus stations

Apple Store Connect


Visit Onda 
in Apple Store

Team Member