Hi. I am Dennis
I design because I love problem solving and discovering reasons behind why people behave the way they do
8+ years of exp in product design as well as 3+ years of experience in product management.
Skilled at tackling ambitious projects by developing design strategies that break down problems into manageable and actionable steps.

Adaptable & flexible
I approach design in the same way I approached competitive gaming
At one point, I was in the 99.5th percentile on the Starcraft 2 North America server. I learned how to tackle ambiguous problems by gathering information on the fly and coming up with quick solutions, all while staying flexible in every game I played.
Project One - Nobul messaging system revamp
Streamlining user communication experience to stop the bleeding and boost the earning
The story of how I leveraged research data to execute a project that revived a heavily underused messaging system, kept users engaged, and helped the company earn thousands more in monthly revenue by improving its ability to keep track of transactions going through its marketplace.
Project overview
A multi-stage overhaul project that lasted 21 weeks
Company & project background
How was the messaging system related to revenue?
Nobul's core business model is to profit off of connecting home buyers or sellers with reliable real estate agents who collaborate with us by taking a small cut off of the total transaction amount when a partner agent has successfully closed a deal with the client we have introduced to them.
To ensure that the company can keep a good track of the transactions happening through its marketplace, a relatively simple agent-client messaging system was introduced during late 2018.
The observed problem
The messaging system became extremely under-utilized and was no longer able to track transactions
Half of the homebuyer/seller user base were either actively asking for or agreeing to the usage of alternative communication methods.
Most of our partnering agents would rather use other tool to talk with their clients after getting hired.
Our task
Stop the bleeding
Between Jan 22 to Jun 22 alone, our finance team found that the number of deals closed behind our back was in the low double digits. As a result, the product team was tasked with finding ways to reduce the number of missed deals by at least 50%.
Secondary research
Taking a deeper look at the issue with a product audit
To figure out why many users weren’t using our messaging system, I did a thorough product review. This included market research, user journey analysis, and a deep dive into the user feedback we’ve collected over the past 2 years.
Problem area No.1
Disjointed communication experience
The messaging system's inability to send photos and documents, along with the lack of simple methods for users to perform essential tasks such as hiring agents or scheduling home tours, frequently forced users to leave the system to complete their tasks.

Problem area No.2
Poor communication efficiency
Feedback from the past couple of years indicates that our agents have been frustrated by the inability to send photos or documents and were also deeply bothered by the lack of features to help streamline their workflow.
Problem area No.3
Visual clutter and poor usability
Many of the users we spoke with over the past years mentioned problems related to visual clutter, unintuitive information presentation, and unthoughtful feature implementation.
Unexpected constraints
Pushback from both the upper level stakeholders and developers
There were two main reasons for why the communication system never included basic functions like sending or receiving photos or documents.

There was a misguided assumption about the benefit of not letting agents sign contracts with clients offline.

The cost of implementing a secured backend system for storing sensitive files was too high.
Problem validation
Additional user research
We interviewed 20 people to validate the user pain points identified during the product review and to narrow down the list of items we should prioritize in this project by determining what mattered most to our users.

Removal of detrimental constraints
The research results, along with finding a cheaper way to implement the file storage system, got us the green light to move forward with our plan
With our earlier findings confirmed and the realization that we could implement the file storage system for a much lower cost, leadership quickly agreed to our suggestion to add file transfer support to the new messaging system.
Define goals
Create a more streamlined and efficient communication experience that encourages users to stay engaged with each other
After discussions with the technical team, we decided to focus our efforts on addressing the most impactful issues affecting the overall user experience.
01
Reduce the chance of users exiting the messaging
02
Make it easier for agents to provide updates to their clients
03
Address usability issues whenever it is technically feasible
Plan of execution
A multi-phase rollout plan that enables us to deliver improvements quickly to alleviate the problem, while also giving the team enough time to address the issue more comprehensively
Remove Point Of Exits
After discussions with the technical team, we decided to focus our efforts on addressing the most impactful issues affecting the overall user experience.
Add automation
Following the successful completion of phase one, I moved on to phase two, focusing on adding automation to the client progress update workflow in a technically feasible way.
Phase one
Create a more immersive experience by enabling users to accomplish more tasks within the messaging system
Offer users with a communication experience that enables users to smoothly transition from one task to another without any hindrances during their real estate journey.
Define scope
Choose the point of exit to be eliminated
Even with a file transfer system, users will still face many exit points during their real estate journey, making it unrealistic to fix them all. After looking over the user journey charts for home buyers and sellers and considering our project timeline, we've picked five main exit points to focus on.
POEs to be eliminated
Problem areas that could be resolved by solutions that are more technically feasible under the time and resource constraint we had.
Revisit in the future
Problem areas that were more challenging to find a solution for the time being. It is likely that we would want to revisit them in the future.
Not within scope
Problem areas that are either too technically difficult to provide a solution for or are already resolved by the file transfer system.
Exploring various approaches
Identify the optimal approach that delivers the desired user experience
We explored a few different approaches, ranging from highly technically feasible solutions to options that required the development of new components with UI interactions that have never existed in our design system before.
Approach 1
New page / Tab
Rejected
Approach 2
Overlay
Not good enough
Approach 3
Side panel
Perfered
Actualize the solution
The 'Quick Action Hub'
In the end, we decided to name the new feature the 'Quick Action Hub.' This feature would be a slide-in side panel that allows users to compare agents, hire them, book new tours, reschedule tours, and leave reviews, all in one convenient place.
Tackle agent compare & hire
Streamline the agent comparison and hiring experience within the messaging system
Typically, users would need to leave the messaging system twice: once to compare profiles of different agents and once to proceed with hiring the ideal agent. In both instances, the user would navigate to the agent's profile page.
The goal was to design a page that allows home buyers and sellers to quickly glance at an agent's background and proceed with hiring if needed.

Tackle home tour scheduling
Provide users with a faster way to schedule their new home tours
Previously, users had to navigate to the property detail page and go through the home tour booking flow there if they wanted to book a property tour with the agent they were working with.
To eliminate the need to leave the messaging system and book the tour elsewhere, I designed a slightly modified home tour booking flow tailored for booking with agents you are currently conversing with.

Tackle home tour rescheduling
Make home tours rescheduling more convenient when using the message system
We observed that users’ rescheduling requests often go unnoticed by their agents, resulting in many users abandoning their agents and seeking new ones elsewhere."
To enhance the tour rescheduling experience and prevent user dropoffs, I designed a mini home tour booking flow specifically tailored for booking with agents directly within the messaging system.

Tackle agent review
Provide users with the ability to leave an agent review directly within the chat UI
Leaving a review has always been such a hassle. The fact that users would have to leave the chat UI and find where they could leave a review for their agent was a huge turn off for many user.
To eliminate the need to leave the messaging system and book the tour elsewhere, I designed a slightly modified home tour booking flow tailored for booking with agents you are currently conversing with.

Layout - before vs after
A more effective utilization of screen real estate
The screen space allocation for the contact list and chat UI was redesigned to provide users with access to more features and functions on a single screen, maximizing screen real estate utilization.
Usability optimization
Main quality of life changes in phase one
Redesigned the search bar for better accessibility.
All contacts now display a portrait photo along with their full names.
Added quick access to the block, delete, and pin-to-top features.
Added a new visual divider to help user tell contacts apart.
Phase one impact
Overall higher user satisfaction
Although the data science team did not observe a noticeable improvement in user metrics during the first month after launching the new messaging system, our customer service team has reported that more users now view the Nobul messaging system more positively than before.
"Using the messaging hub right now make much more sense compared to before."
From a home buyer in Toronto
"I no longer feel like I have to provide personal contact methods when I am deciding on who I want to work with!"
From a home seller in Edmonton
Phase two
Give our agents more reason to communicate through Nobul actively
Incentivize agents to actively use our messaging system by introducing a new feature that speeds up the communication process.
Define an agent's journey
Draft up the home buying and selling workflow
I drafted up a chart to visualize the different stages that agents would need to go through during a typical home buying and selling journey.
Pre-generated messages
Create a versatile template that agents can leverage at every stage of the process
After analyzing user interview feedback and reviewing conversation samples from the data science team, we concluded that providing agents with pre-generated messages for their updates would be the most impactful feature to enhance communication efficiency.
After extensive exploration, we have finalized the format presented below.

Usability optimization
Notable UI optimization done in phase two
Gave the property card a redesign for a cleaner, more modern look.
Provided users with easier access to voice and video chat in the mobile UI.
Refined the message timestamp display format to enhance readability.
Added a new visual divider so users can easily tell messages from different days apart.
Project results
Surpassed all expectation
After approximately six months of observation, I was thrilled to discover that the project had resulted in a significant improvement in all key user metrics, and we also observed a considerable decrease in the number of missed deals.
Home buyer & seller retention
+75%
From 50% to 88%
User engagement post hire
+248%
From 25% to 87%
Transaction tracking efficiency
+73%
From low 2 digits to mid-single digit
Project Two - VPP home rental listing workflow redesign
Restore user trust in a declining rental marketplace by creating a more convenient and efficient workflow for listing rental properties
The story of how I significantly enhanced the credibility of rental properties listed on the largest Chinese-facing home rental marketplace in BC by offering users a property listing process that allows them to input information more efficiently.
Project overview
Part one of the multi-stage revamp of VanPeople classified listing home rental experience
Company & project background
A once-thriving marketplace was experiencing a rapid decline
Ten years after its launch, VanPeople's classified listing platform has thrived, thanks to the massive influx of Chinese international students coming to study in Canada.
However, as the user base has grown younger, VPP's outdated design, lack of modern filtering features, and the abundance of low-credibility rental listings have made it difficult to attract new users.
The observed problem
More people have started to search for rental properties elsewhere
By comparing the data collected from December 2019 to June 2020 with that of December 2018 to June 2019, we noticed troubling drops in some key metrics for the home rental section.
The home rental section's average daily visit count fell from 12.5k down to 9.7k
The average amount of new rental listings being posted each day went from 130 down to 118.
How we responded to the problem
Covid? No! Actually the issue we had at hand was real bad!
Initially, we assumed that the decrease in activity on our platform was due to the impact of COVID-19. However, we noticed that even after the lockdown had ended, both the number of daily visitors and new listings continued to decrease over several months. This led us to conclude that the underlying cause of the problem was likely something other than the pandemic.
Secondary research
Find out more on why the numbers were dropping left and right
To gain insight into why the platform was losing popularity, we initiated a comprehensive research effort, comprising surveys, 1 on 1 interviews, user journey mapping and market competitor analysis.
Notable findings
Significant issues were identified regarding listing credibility, search efficiency, and the property listing user flow
Major trust issue
17/20 of the participants complained that many rental listings did not seem that credible.
Difficult to use
16/20 of the participants reported that looking for homes was taking too much effort as there were no ways for users to narrow down their options.
No incentive to contribute
4/5 of the participants who have posted ads before reported that they do not really feel like adding detailed rental info when they list properties on our platform.
Research insights
The lack of incentive to add information to the rental listings, resulting in poor listing credibility, created a chain reaction that continually drove users away
The need to manually input all information was discouraging users from adding detailed rental information to their listings
The lack of detailed information made the properties listed on VanPeople appear untrustworthy
We were also unable to add meaningful property filters because we lacked the detailed information necessary for filtering
In the end, users were losing trust in our platform because nearly all of the listings on VanPeople were lacking credibility
Where do we begin?
Rebuild user trust by first revamping the property listing workflow
By streamlining the property listing workflow to encourage users to create detailed and informative ads, leading to an increase in high-quality listings.
As more listings look more credible because they include more information, it’ll kick off a positive chain reaction that encourages more users to engage with the platform.
Allow users to list their properties for rent faster than before
Encourages users to start providing more detailed property info
New design foundation
The new design was developed based on a style guide that ensures greater design consistency, as inconsistency can undermine user trust
Notable constraints
Unable to efficiently implement progressive UI changes due to existing users' attachment to the current setup
The majority of VanPeople users were older individuals who were less adaptable to changes and had become accustomed to the platform's consistent appearance over the past four years.
Thus, drastically changing the look and feel of everything all at once was deemed too risky, as it might create too much disruption to the current user experience.

Design requirements
Compiled a list of essential elements that were most important to our users when listing a property for rent
I identified the essential information necessary for creating a comprehensive listing post through meticulously reviewing the contents of over 100 listings.
To validate the list and ensure it aligned with the actual expectations of our users, I consulted with 5 colleagues who had actively used our rental service in the past few months.
New enhancements
Introduced new interactive components to improve workflow efficiency
To streamline the property listing process and reduce user time investment, I incorporated elements such as checkboxes, dropdown buttons, and auto-complete mechanisms into the design. This allowed users to avoid manual entry of detailed property information.
Version 1.0
An improved one-page design basing off of the old design
Due to time constraints, I opted for a straightforward one-page design, essentially an extended version of the previous layout.
Following a brief validation test with the same five colleagues I consulted earlier, all five agreed that they would be more likely to input detailed property information using the new design.
Problem with V1.0
Seeing too much information when I don't need to see them
Despite these positive findings, the initial draft was considered overwhelming by the participants. One out of the five did not complete the test, citing that the page required too much information at once.
It became evident that consolidating too much information into a single long page caused information overload.
A new approach
If users felt like they were seeing too much at once, how about breaking the whole process into smaller chunks?
I created a three-page, multi-stage workflow designed to guide users through smaller chunks of the process, one step at a time, with some pages being potentially 100% skippable.
V2.0 - The final design
A three-step process that enables users to list properties for rent more quickly and with more detailed information than before
I conducted a validation test with my co-workers to verify that this new approach was more effective at conveying information than the initial version and the consensus among participants was that the redesigned interface significantly improved usability and information digestibility.
Step one
Fill in basic listing information
I updated the system so that users can no longer leave the rent field blank, eliminating the option for people to list 'price is negotiable in person' as the pricing for their listings.
In addition, to encourage users to upload more photos when listing their properties, I prioritized the "upload photo" feature as the first thing users see when landing on the page.
Step two
Fill in additional detailed info
The second step was designed for users to swiftly input all property details with just a few taps.
To streamline the workflow's overall user experience, the majority of non-skippable fields are positioned at the top for easy visibility at first glance.
Step three
Fill in contact info
This final step allows users to quickly input their contact information. All details are now auto-filled if the user has an account with us.
I have included the option for users to withhold their detailed addresses to avoid potential harassment. Additionally, users now have the option to hide their email from the public to minimize solicitation.
Additional design validation
The new workflow was 3x faster than the old
A final round of A/B testing were conducted with some of my co-workers after the new workflow went live on our beta testing server. Participants were tasked with listing a property with detailed information using both the old and new processes on separate occasions.

Avg. property listing time
-5 mins
From 7 mins to 2 mins
Result of the trust rebuilding effort
Notable improvement in both listing quality and quantity
Two months after rolling out the new property listing flow and updated detail page, we've observed a positive trend: an increase in the average number of daily rental listings. More importantly, these listings now feature much more detailed information.
% of listing with photos and a clear rental price
Increased by
+200%
From around 3 in 10 to 9 in 10
Avg. number of rental listing posted daily
Increased by
+22%
From around 110 to around 135