“I believe that the race to success is not speed but stamina.”
Yeng Bun is involved in analyzing, modeling and programming real-estate data. He is the Principal and the Co-Founder of Nubram,LLC- a family business founded in order to integrate analytics with his hands-on experience to guide asset management and investment in affordable high-quality urban housing markets. Previously, he worked with Zillow.Inc (https://www.zillow.com/) and worked on many different analytics such as Zestimates, Rent Zestimates, Home Price Index and Rent Index. He has several patents to his name too.
Q1. Describe your life journey. What did you do at Zillow for 12 years?
Before Zillow, I enjoyed working for Wall Street firms modeling and programming with high frequency trading data. Unfortunately, working remotely for firms in New York did not work out for me and my family. So, I decided to gamble working for a startup-Zillow. The company grew fast, and I had so much work to do. My boss often told stories of my “daring exploits keeping all the balls in the air almost single-handedly”. I had my hands on many different analytics, but a few that in my hands the longest were Zestimates, Rent Zestimates, Home Price Index and Rent Index.
Q2. You were involved in the pioneering of Zillow home valuations. What were the main challenges you faced in building the Zestimate Models?
The main challenges include the following:
- Missing data: No transaction in some regions. For example, transactions are not public records in non-disclosure states in the US such as Texas and Utah. Another example is missing attributes such as home size, number of bedrooms and bathrooms.
- Sparsity: Only small transactional dataset available in some regions or market segments. For example, rural counties may only have a few transactions per month. And there are always few transactions for high-end and low-end homes in many neighborhoods.
- Dirty data: Incorrect or out-of-date. Who should I believe when there is disagreement between public records and homeowners? Owners claimed to have upgraded and added more living space, but I have no way to verify the info. How do I know if a homeowner just wants to manipulate the Zestimate? In fact, there was a feature for users to instantaneously see changes to Zestimates with changes to home attributes. Another example of dirty data occurs when the sale price is low because of transactions between relatives. If a father sells his house to a child, I may be able to filter it out based on their last names. But what about a grandmother selling to a granddaughter? Also, sometimes, a low sale price can be the fair price because the home is distressed, but it is not easy to verify it without physically inspecting it.
Q3. What do people misunderstand about their Zestimate valuation?
A common misunderstanding is the purpose of Zestimates. Zillow has always been upfront and clear about the use of Zestimate as a starting point. But many people repurpose it to be the final point. Every quarter, I carefully calculated and published Zestimates accuracies. This helped mitigate loss of trust caused by this misunderstanding. Also, over time, Zestimates’ accuracy improved as Zillow continued to prioritize its investment to improve Zestimates.
Q4. You have several patents to your name. Can you share some thoughts about them?
I feel very privileged to be able to work on projects that were innovative and patentable.
Q5. What motivated you to start Nubram?
I wanted to integrate analytics with hands-on experience to guide asset management and investment in affordable high-quality urban housing markets. I knew some analytics in the real estate space, and I had some hands-on experience in buying, fixing, renting, and selling houses. This was fun and profitable.
Q6. Please describe how Nubram could help a potential client?
Nubram is a family business that does not serve any outside clients now. Hopefully, when my kids take over this business, they can expand it to include serving outside clients.
Q7. What moment changed your career trajectory forever?
I started my career doing basic research with fast high-flying objects such as hypersonic jets and superguns. These guns could shoot projectiles to the moon if we could control their trajectories. One day, while I was excited to show off my tools for designing such projectiles, the sponsor was excited about shooting chickens out of these superguns. It was about testing collisions between flying birds and slower flying objects at low altitudes for research in material science. In other words, there was going to be a shrinking of funding for basic research in my areas of interest. That moment changed my career trajectory forever. I then decided to use my knowledge and experience in other areas centering around statistical computing and data mining.
Q8. What headline will the WSJ write about you in 2025?
All 3 of my children entered college early and graduated from University of Washington by the age of 19 without attending High School. Ironically, I believe that the race to success is not speed but stamina. I liked reading the story of Rabbit and the Turtle as a bed-time story to my children when they were young. And they fell asleep fast; it must have been a very boring story to them. It turns out that they have a lot of stamina with resiliency and can overcome challenges to be successful in their own careers. And I am proud of them growing up to be good people.
Q9. What would be the most important thing you’ve learnt from a peer and how have you used that lesson in your day-to-day life?
Trust! One of my partners used an analogy of a basketball team. As a point guard, when he passed the ball to me, he trusted that I would take care of the ball or score points for our team. For my day-to-day life, I trust my partners/co-workers to do their part as they trust me to do mine for our joint projects. With trust as a key ingredient, we can succeed in project after project.
Q10. What is your opinion/outlook regarding home prices over the long term?
Long-term, people will continue to migrate to some urban areas. Covid-19 acts as a break to slow down this trend but it is temporary. The more population growth in a region, the more home prices will rise in that region. The prices will also strongly depend on the costs of land, building material and labors. While inflation may push up the costs of building material and labors, technologies may enable efficiencies that reduce these costs. However, land costs will go up because of its scarcity. As Mark Twain had said “Buy land, they’re not making it anymore.”
For more such insights, http://www.home.llc