Chun San Yong

Improving the confirmation screen

How I reduced cancelled investments by 60.1%
 

Context

Endowus’s core feature is a goal creation feature, where users set financial goals that they want to achieve through investing. For new users, this is the first thing they go through in onboarding.
 
 
The onboarding flow: goal creation, then personal information, then confirmation, then waiting for account opening.
The onboarding flow: goal creation, then personal information, then confirmation, then waiting for account opening.
 
Problem
23% of new users were calling or emailing us to cancel their initial investments.
 
Goal
Reduce the number of cancelled investments, with minimal changes to the screen’s content structure.
 

What I did

 
Frame the problem of unfamiliar language
The immediate problem area was pretty obvious: a question on the confirmation step that asked users whether they wanted to invest or not.
The odd and unfamiliar language of “create investment request” confused users, who would select “Yes” even though they had no intention to invest right away.
 
 
The unfamiliar language made users select “Yes” or “Now” even though they had no intention to invest right away.
The unfamiliar language made users select “Yes” or “Now” even though they had no intention to invest right away.
 
Balance precision and simplicity with stakeholders
Nailing the copy was a challenge. I had a few short lines to explain the complicated consequences of each option and our multi-step transaction process.
In the end, I prioritized simple and familiar language over precision. This was a pretty controversial decision!
 
Craft content patterns for variants
I had to deal with conditional variants and legacy design variants. To wrangle these variants, I crafted and stress-tested modular content patterns that could work across flows and conditions.
 
 
 

Final design of v1

 
Before and after on our web app.
Before and after on our web app.
 
Before and after on our mobile app, with new content patterns included.
Before and after on our mobile app, with new content patterns included.

Impact of v1

Investment cancellations fell by 60.1% after a month, from 24.3% of all investments down to 9.7%.
 
 
 

Iterating towards v2

2 months after shipping v1, we started on v2. We’d made good progress in clarifying the options and consequences of when to invest, but v1 had its own issues.
 
Problems
Our mobile users:
  • didn’t understand what each option meant because of the chip interaction pattern
  • dropped off on the confirmation screen because they felt intimidated by the decision of when to invest
 
Goal
Make it even clearer and give mobile users confidence to move forward on the confirmation step.
 

What I did

 
Frame the problem, again
To figure out the content requirements, I audited the v1 content patterns with fresh eyes and mapped the different scenarios.
 
Mapping the v1 content patterns and sketching quickly to think out loud.
Mapping the v1 content patterns and sketching quickly to think out loud.
 
As I mapped the scenarios with my product designer, it became clear that:
  • the question at hand was really about capturing user intent, and it was irrelevant to tell users about next steps.
  • new users faced a big time lag between the moment of confirmation and the actual investment. We had an opportunity to give users more confidence by addressing this time lag upfront.
 
 
When in doubt about conditional variants, make a matrix.
When in doubt about conditional variants, make a matrix.
 
Define a content approach
With this new understanding of the problem, my approach was to:
  • Keep it simple to lower cognitive load on a crowded confirmation screen.
  • Reassure users that we would help them succeed with next steps via reminders.
  • Present sharp binary options to clarify a user’s intent, with a focus on explaining the “Invest now” option.
 
 

Final design

 
In v2, I also tidied up the page’s content hierarchy and narrative to give the confirmation step its due sense of being an ending.
In v2, I also tidied up the page’s content hierarchy and narrative to give the confirmation step its due sense of being an ending.
 

Impact

No data yet, v2 is currently in development.
 

Reflections

 
Tiny copy changes help a team learn things and build momentum
I saw many content problems to fix when assessing v1 and advocated solving them, but the extremely contained scope gave useful data and proved the ROI of content. It helped me build momentum and trust with stakeholders, leading to a bigger appetite for improving content in v2.
 
 
 
badge