Not every creator is interested in monetizing their direct messages. Therefore, we need to narrow down to a group of creators who is intersted in monetizing their direct messages and target them through our product and promotion.
Twitter Spaces
Clubhouse
Medium
Freelancer.com
Fiverr
Youtube
Tiktok
After the business team got the quantitative data from surveys, I conducted user interviews to dig deeper into potential user’s motivation, painpoints, and goals. I was able to connect with several creators from different platforms including Instagram, Medium, Twitter Space, and Clubhouse. In the end, we decided to make 5 variants including Fiverr.
“Sometime when I gave them mentoring advices on their career, I wish I could be compensated”
“I have always wanted to find a way to monetize my audiences”
Before each session, I would write a usability test script to help me stay on track when conducting the test. During the sessions, I would make it clear that we are testing our prototype rather than them. After the test is finished, I would use infinity mapping to help me find some comman patterns in the test. In the following example, I am showing the changes I made from the initial state of the application before I arrived.
1: 90% Users don’t know this screen is their inbox when first landed in the app.
2: The engineering team puts the buttons there wishing that users would click them because they are obvious; however, during the test, users are all confused about what buttons are and had to read the description. This is because users don’t expect the buttons to be there.
This screen is one of the hardest screen to fix, that we had to go through multiple iterations to arrive this final result. As you can see, the end result looks nothign like the initial screen. I was also super confused when I first see this screen. After talking with the engineering team, I began to understand that the core function of this screen is to enable or disable prepay. There were so many things wrong in the initial design. They were trying to incorporate everything in the screen, but it just makes the screen too crowded and may cause cognitive overload. In many cases, when a user sees something like this, they would simply give up.
Every friday, we meet on discord to show our progress. We will also discuss the next move of the app. One of the things that I reallyt like about this startup is that everyone’s voice matters.
When the internship was almost over, my project manager approached me and said “If I am going to do it again, it would just be you and me for the first 2 months. After we are done with design and testing, I am going to hire the engineering team.” I absolutely agreed with him. I think that would save everyone so much time, because we are not building stuff over and over again.
I think simplicity is very important, so I choose a very minimal approch on the UI color pallete. As a result, the actual content can draw people’s attention instead of UI elements.
.gif)
In other apps available today, you can only split evenly with your friends. However, very often, people want to edit individual’s amount in the end. Nova Pay allows you to manually input the amount you want to request from a person, and the app will intelligently calculate the rest.
Instead of asking the user to remember how every bill needs to split and manually calculate the amount, I designed a system that users can simply choose from a list of payments and the app calculates the amount for them.
.gif)
Simply scan the QR code on the desktop and complete your payment on your phone within seconds.
Week 1-2
.png)
Week 3-6
.png)
Week 7-9
.png)
“Remind others to pay me back feels awkward”
Tony, 21
Undergraduate student
“I am more used to the older methods, such as checks and deposits.”
Bethany, 40, Neurologist
Introducing recommendation prioritization empowers analysts to efficiently manage their tasks by focusing on high-impact recommendations, improving workflow and inventory management decision-making.
Information is presented precisely when it's needed. It provides a concise AI Recommendation summary, critical AI alerts, and a detailed AI recommendation page with evidence for informed decisions.
The AI transparency improvement enhances user decision-making by providing full transparency. This includes AI Alerts that offer clear insights and a redesigned recommendation detail page for easier understanding and informed choices.
Recommendation Acceptance Rate
Time to Process Recommendations
Painpoint 1: Hard to collaborate with AI
Analysts feel it is hard to process AI recommendations, because all recommendations are presented in a flat way without prioritization
Painpoint 2: Information presented not promotive
In this modal, an anlayst is about to make a decision on 1500 recommendations, but the modal doesn't provide enough information to help the analyst make the decision.
Painpoint 3: AI Evidence not transparent
It is hard to process the AI Evidence of recommendations becuase of confusing data visualization and excessive use data science language rather than user language.
User language instead of DS Language
Use language that is familiar and easily understandable to the user, rather than technical data science jargon, to ensure better comprehension and user-friendliness
Only show relevant information
Display only information that is pertinent and valuable to the user's current context or task, avoiding the clutter of irrelevant data to enhance user experience and efficiency
Align with user’s mental model
Design interfaces and interactions in a way that aligns with the user’s intuitive understanding and expectations, making the system more user-friendly and easier to navigate
Through this redesign effort, we are able to increase our Recommendation Acceptance Rate by 50%, and decrease min to process each recommendationexpand from 3 facilities to 18 facilities.
Recommendation Acceptance Rate
Time to Process Recommendations
.gif)
.gif)
