Improving Teambition’s App Retention

I led the redesign of the most visited pages to improve app retention. I utilized strategy frameworks, defined key metrics, discovered core actions, pitched the ideas, and shipped a polished design.


Teambition is a Chinese project management platform that makes it easier for teams to track their work with efficiency. Some similar products are Trello, Basecamp and Asana.

Starting late last summer, I received an ambitious project to improve the mobile app retention. As you might imagine, breaking down this daunting goal can pose some unique challenges.

The new design will be launched in late November.

new design


I came across three main challenges:

1. Fuzzy Problem Space

Before this project, I had no idea how to improve mobile app retention. So I read a bunch of growth articles. Specifically, I drew inspiration from Andrew Chen and Sarah Tavel.

  • I learned from Andrew Chen that to improve mobile app retention, I’d better focus on the first-time user experience.
  • Sarah Tavel’s The Hierarchy of Engagement also blew me away. With the help of data scientists, I used her growth framework to chunk the complicated project down with four core actions and did tons of iterations to make each piece really simple to use.

2. How to Convince CEO

Teambition is a nimble startup, so I had a great opportunity to continuously persuade CEO why it’s a problem and how to solve it. As a design intern, it was not an easy task, so I took a two-step approach:

  • First, I did a competitive analysis and pointed out several horrendous pitfalls during new users’ onboarding process. He agreed, so I made sure we were on the same page.
  • Second, based on his feedback, I brainstormed several ideas and invited him to pick some that piqued his interest. It worked. It not only made me sound more reliable but more importantly kept the project rolling with clear directions.

3. Quick and dirty Research Methods

Traditional user research methods such as interviews and user testings are not valued enough here. I guess it’s because the designer-centered culture encourages more of heuristic evaluations and internal dogfoodings. I used some quick and dirty research methods as workarounds. I found three methods extremely helpful to gather insights:

  • Competitive analysis
  • Data analysis
  • Customer complaints

Scroll down to learn more about how I integrated research into design process.

data analysis
data analysis

First-time User Journey

"...where I see that most of the leverage in improving these retention curves happen in how the product is described, the onboarding flow, and what triggers you set up to drive ongoing retention." - Andrew Chen

With the first-time user experience in mind, I focused on three key touchpoints for new users to discover our product value as soon as possible:

  • Onboarding experience
  • Creating a task
  • Task details page

1. New Onboarding: Create projects at your fingertips

One area I wanted to improve was onboarding. I wanted new users to create projects and take advantage of benefits to track the progress. But even creating projects is burdensome for new users.

So in the new design, I encouraged users to pick some templates out of six most popular project templates. I also added a transition from creating a project to creating a task so users don’t need to wonder what the next step should be.

new onboarding

2. Creating a Task: Delightful than ever

Task is the most important piece to our product. The existing design was cluttered and users felt like they had to enter a lot of stuff.

In the new design, I highlighted one essential piece, title, with more white space in addition to de-emphasizing other less important text fields.

creating task before
creating task after

3. Task Details Page: More Effective Call-to-Actions

For a project management platform, task is everything.

"What matters is not growth of users. It's growth of users completing the core action." - Sarah Tavel

I unpacked the concept of retention based on Sarah Tavel’s engagement framework and narrowed down to four core actions I would encourage new users to do:

Sharing a task
Sharing a task
Changing task stage
Changing task stage
Inviting colleague
Inviting colleagues
Commenting on a task
Commenting on a task

I partnered with our data scientist and used data to justify those four core actions. Users who did those four core actions have significantly higher retention rate than users who didn’t.

task page before
task page after

Design Details of Task Page

I’d like to talk more about Task Details Page in detail. Not only because this is the one I put the most effort into, but also because I learned how to design under constrainsts.

1. Sharing a Task: Untraditional research helps

sharing before

During heuristic evaluation, it was confusing even for me to figure out what “Read Only” mode was and what differences were between several modes of comments.

Data also validated that. Only 16.7% people shared successfully after they hit the sharing button on mobile. Something must be wrong with the sharing.

I did several rounds of iterations. The first iteration mainly focused on the wording. I also changed the order of sharing methods based on relative usage frequency.

Obviously, that was far from enough. Not until I discussed with customer success manager for customer complaints did I discover the tradeoff between privacy v.s. convenience.

  • Privacy: Users are hesitate to toggle the “Read Only” button on (i.e. visitors don't need to log in to see the content) because they worry about information leakage.
  • Convenience: Users need to share tasks to colleagues on Wechat.
    (One interesting side note, Chinese people are so accustomed to Wechat that they will use it for everything, including work. They even bother to download our native app.)
sharing after

This insight opened a door of new possibilities. The new design allowed users to share without toggling the “Read Only” button on, i.e. they can share to their colleagues internally or externally under privacy mode.

It also mirrored the existing capability of web users. If users can copy a URL from a browser and send the link to their colleagues through Wechat, why can’t they do the same thing on mobile?

2. Changing Task Stage: Use data to justify intuition

Task is the core piece of our product. Project management without changing task stage, say, from “To Do” to “Work in Progress”, is impossible.

Sounds obvious, but I couldn’t get the engineers to buy in. So I approached our data scientist to ask for more evidence. It worked. Data shows that the average number of events per action is as high as 20 times, much higher than other events such as copying or archiving.

data analysis moving
Average number of events per action (Changing Task Stage) is as high is 20 times.

As you can see, one of the most fascinating thing to work in a in-house design team is the ability to work closely with the engineers and convince them of your design decisions.

changing task stage after

I opted for more obvious call-to-action by moving the button to the upper right corner.

I also reduced the stepes of changing task stages by landing users on the task page page without touching project or task group. This, of course, can save them a lot of headache.

3. Two Other Core Actions: Inspirations from design trends

Both inviting colleagues and commenting on a task are low hanging fruits. The goal is to make it more obvious to invite colleagues and and make it easier to engage in conversations. I drew inspiration from Dropbox’s header with circle images and Facebook’s new messaging design paradigms separately.

The header with circle images is a more effective call-to-action to encourage people to invite colleagues than a grey cell hidden below the main task

explorations of header
Final Version
final version of header

The messaging design paradigms empower people to converse better than message board styles.

comments before
comments after

4. Potential Reading Area: Killed due to engineering constraint

Engineering constraints are everywhere. Some of them could only be discovered with the help of engineers.

One area that I wanted to improve was how people read the task on the page. This insight was based on the observation that people would read the title combined with the notes. The first iteration was to separate the elements with two groups: reading area for immersive reading; action area for quick look.

reading area before
reading area iteration
reading area after

However, this idea was killed due to engineering constraint. In short, since we already empower users to change the order of fields based on their own needs, we can’t hard code to change the order (such as placing notes right after title) any more.

Good news was we still managed to make the task easier to read by improving visual hierarchy and space contrast.

5. Other Changes: Every detail matters

Aside from those core actions, I also did a variety of design treatments to find opportunities on improving user experience on task page:

subtask after

Helped people break up the work by expanding subtasks in a checklist format.

calendar after

Made the date easier to select by changing from default date picker to a calendar view.

scrolling after

Made the new activities easier to read by scrolling automatically to the bottom of the task page from their inbox.

gesture after

Improved navigational gestures by enabling people to swipe back out into Task Panel.

photo uploading after

Provided a more convenient photo uploading experience by placing recent photos right at the top of menu.

What’s Next

I will validate all the design decisions based on the performance of key metrics such as active users, number of events and retention rates.

The biggest lesson I learned in this internship is the tradeoff between quality and cost. Obviously nothing is ever “done” because there’s always another idea. However, designers have to learn to balance between time, cost and quality constraints.

Designing Teambition mobile app has been an incredible adventure so far, and I’m really grateful that I got to work on so many things, and that Teambition offered me a project with a big impact that I would have ownership of.

Read Next

Redesigning AI training platform

Enter Password