Coffee Meets Bagel
A dating app that helps people find their special someone.

Available on iOS and Android

3 cmb screens.jpg

How Coffee Meets Bagel works

Everyday at noon, users receive several, highly curated matches based on our sophisticated matching algorithm and can choose to Like, Pass, or Connect. If there is a mutual Like, a private chat will open for both parties to chat for seven days.

What is my goal

My goal is to design an engaging experience that connects people virtually and transforms the virtual connection into something real. There are certain things in real life dating that cannot be replaced by online dating. The differences often lead to user frustration. To reduce this frustration and create something better, I listen to my users' needs, and design for their need instead of their want.

My role

I design features from conceptualization to launch. I follow an iterative process that incorporates ideation, information architecture, user flows, mockups, prototypes, usability testing, and team feedback to create meaningful experiences and help people meeting their special someone.

Who do we design for

When designing a dating app, we always have to keep in mind that

1. Men and women behave very differently when it comes to dating online

2. Other than gender, differences in age, location, popularity varies each person's experience

3. The ultimate goal that motivates users is to connect with people


There are other things that make CMB uniquely different from other dating apps. We value quality over quantity. We optimize for the female experience. These core values are the foundation for developing our products.


The way we build

  • Frequent and fast feedback loop. In the early stage of a product feature, once we have an initial goal, we design low fidelity prototypes, send out surveys, and look at supportive data to help us define the product. We iterate on the idea and test with users that have different levels of familiarity with the app and different genders. Depending on the product type, we may have more criteria of the testers.

In early stage of product development, make prototypes and test relentlessly to make iterations before defining the product will help us verify assumption and adjust directions.

  • Collaboration and openness. We share initial designs with the team as early as we can. This helps seed our design culture and the feedback from the team is so valuable that it always reminds me that I have to listen to the users more.

  • Good process makes good product. Rethink what works and what doesn't. Having a good process will minimize time investment and mistakes from communication gaps. When designing an innovative idea it’s possible that a different process should be utilized because the existing process doesn’t fit the need.

Features I have worked on ↓

Coffee Meets Bagel - Skip the Line
Get seen faster by your matches


Popular users have a long line of matches who are waiting to be shown to them. When a user likes a popular person, it's difficult to stand out and it will take days to show up to the popular user. Both parties are missing the opportunity to make a meaningful connection.


An existing feature called Woo attempted to solve this issue by allowing users to send a virtual gift to express their special interest. It also allowed users to jump to the front of the line to be seen faster by their matches. But the adoption rate is low. And it caused a lot of user confusion.

Group 1 (2).png

What customers are saying

"What is Woo?"

"Why can't I get connected with my matches. I wish there's more I can do."

"I'm receiving many profiles but are they just liking everyone?"

Key Insight

This is the conversion funnel. Users start from downloading until they go on dates. A key insight we learned is that as more time passes between when a user likes a match until when they connect, the less likely they will exchange chats. We decided to create a goal to improve the conversion rate and reduce the time from like to connect so more people will go on dates.

Lily chou presentation.png


User: An easy way to show special interest in somebody when they receive a new match.

Product: Reduce the time from match to chat. Improve Woo's adoption rate.

Research and Prototyping 

Group 6.png

Initially, to improve the adoption of Woo, we A/B tested UI and messaging adjustments. But it didn't make much improvement. We decided to evaluate Woo. We went back to consider our dating app's key questions that inform our algorithms and UX:

1. Which matches will a user receive

2. How many matches will a user receive 

3. When will users receive a match

4. How will users receive a match

Woo is a powerful feature because it allows users to alter two main aspects (#3, #4) of the matching experience. It enables users to send a virtual gift which will alter how their match is presented. And, it accelerates showing the match to the other user. However, the main action of sending a virtual gift was confusing and it didn't communicate to the user that they were moving to the front of the line.

I decided to step back to reexamine the core user problem more than only adjusting UI and messaging.

- HMW reduce the friction of Woo?

- HMW lower the learning curve of Woo? 

Hypothesis 1: Users prefer CMB prompt them when to skip the line 

Hypothesis 2 : Remove the language Woo from user end will allow clearer concept delivery

We decided to switch the value proposition to primarily emphasize skipping the line instead of wooing the other user with a virtual gift.

I created prototypes and gathered qualitative data from both male and female users to validate our hypothesis and produced several iterations. 


- We removed Woo from user end.

- Only prompt a message when users like a popular person. 

- Clearly communicate the benefit of sending flowers.

Group 54 (1).png

Skip the Line has successfully reached the adoption and revenue goal.


We should always make user do less work and do the work on behalf of user. For messaging, we can be more direct and don’t be afraid to be transparent with user.

Coffee Meets Bagel - Activity Report
Reveal user activity data to prevent ghosting


On your match's profile, you can see a dashboard showing how often they initiate chat, how responsive they are to chats, how quickly they respond to messages, and if they have been online recently.


Key objectives

Give users more insight into their matches before connecting. Increase chat reciprocation by making users aware that their activity level is visible to their potential matches.

Research on behaviors, needs, and goals

We made low-fidelity prototypes to validate our hypothesis and set goals to discover the following:

1. How men and women react differently to this concept?

2. Is this data valuable to users?

3. What data is more valuable to users? Why?

We found that both men and woman would find these data points helpful and women were more likely to screen people if they knew their match’s activity level.

How to display the report

We arranged the data in different formats and structures to understand which is easier to understand, and which is more impactful. We validated the design by sending out surveys.

Questions we want to find out in usability tests

1. What is the discoverability of the feature?

2. Is the value proposition clear to users?

3. Is the purchase flow frictionless?


Engadget  Mashable

Coffee Meets Bagel - Photo Lab
Get feedback on finding the best profile photo


There are two parts of the feature: uploading and voting. In the upload stage, users can upload their pictures to get feedback. Once a user has uploaded her photos, other users can vote for the photos that would be good to use.

Key objectives

A lot of users have a hard time picking profile pictures. We wanted to provide a service that allows users to get feedback on their photos.

Ideation and iteration


Use 'likes' on photo as feedback to indicate how popular is the picture


The idea of liking a photo and liking a profile is hard to differentiate. May cause confusion.


Can the problem be solved by UI? Changing 'likes' to 'thumb up'.


Through user testing, we found that the confusion can not be solved simply by UI. We decided to move it away from user's profile and make a separate section. 


Simplify the interaction by reducing the number of photos to 2. 


Showing photos horizontally is a lot easier to compare than stacking photos vertically.

Purchase Photo Report flow

Coffee Meets Bagel - Subscription Design
Introduce premium subscription package to users


Users can subscribe to premium for a given duration. After subscribing to premium, users will be offered upsized Bean packages in the shop.

Key objectives

Design a subscription interface that is effortless and transparent.

Value proposition (saving)

Multiple terms

Keep it short & simple

Call to action

Value proposition

(package description)

Coffee Meets Bagel - Reopen Chat
Reopen chats after it expired in 7 days
Coffee Meets Bagel - Spinner
Spinner animation