As the CTO of Wayfair, Fiona Tan leads the global technology organization responsible for shaping market-leading experiences for the company’s 30 million plus customers, 14,000 suppliers and thousands of employees.
Prior to Wayfair, Tan served as the Senior Vice President of technology for Walmart’s US operations. During her tenure at Walmart, she found herself drawn to the challenge of developing technology solutions for the home category.
“Walmart does a great job of selling everything to everybody,” Tan said. “But I found that the home category posed the toughest challenges when it came to selling goods online, which also made it the most interesting. This was one of the major factors behind my decision to join Wayfair. ”
Tan was also drawn to the prospect of being able to help transform Wayfair at a time when the company has over 30 million customers with $14 billion in revenue, and is poised for what she calls “hyper-growth.”
“A large part of my role at Walmart was focused on developing systems that would help us compete effectively against e-commerce companies, and specifically, Amazon,” says Tan. “I really enjoyed my work there because I had the opportunity to deliver technology solutions at scale during a transformational time as Walmart pivoted to integrate their stores and e-commerce offerings. Today, at Wayfair, I feel like I have the same opportunity to fundamentally transform how we build new things.”
Tan views being able to draw on advances in machine learning and artificial intelligence as being integral to helping the company achieve hyper-growth. In this interview, Tan spoke about her background, how science is helping shape the next generation of customer experiences at Wayfair, and the scientific innovations she finds exciting.
Q. Can you think back to a formative moment that shaped your interest in science and technology?
I was born in Malaysia, and my family moved to Singapore when I was a child. The educational system in Singapore places a high emphasis on excelling in math, science and technology. So it wasn’t just one moment – rather it was the entire educational culture that shaped my interest in pursuing a career in technology. That said, I do recall being drawn generally to the puzzle solving nature of Math and Science, and then more specifically to optimizing the solutions.
I applied to colleges in America for my bachelor’s degree, and remember experiencing an “aha” moment during an introductory class to Computer Science at MIT. During the course, I saw in a very real way how I could actually pursue a career where I got to solve puzzles, and actually get paid for doing it.
At the time I graduated in 1991, many of the thought leaders in computer science were getting interested in how the promise of artificial intelligence could be realized by drawing on techniques from diverse fields like game theory, stochastic modeling and deep neural networks. I took a class on genetic algorithms to study how the principles of natural selection can be used to solve optimization problems. My senior thesis at MIT focused on developing differential diagnostic systems for heart failure, also drew on mathematical techniques used in machine learning.
In the course of my career, my work at companies like Oracle, TIBCO, SAP and Walmart has focused on building large distributed and scalable systems. More recently, at Walmart, and especially at Wayfair, I’ve found myself increasingly getting involved with machine learning and artificial intelligence. I’ve come back to many of the principles I was exposed to during my education. However, we didn’t have the computing power and data back then, and I increasingly find myself nowadays looking at these technologies through the lens of how they can be used to solve everyday problems.
Q. What are some everyday problems that science helps solve at Wayfair?
There’s an incredible amount of science helping shape every aspect of the customer experience at Wayfair. I’ll give you one example.
When you go to our website, and search for a product, we leverage state-of-the-art machine learning techniques to understand the intent behind your search, understand the many nuanced attributes of the products that we sell, and then “match make” your intent to our vast catalog.
Sure, there are many companies that have developed a variety of approaches tackling this problem. However, many of these solutions aren’t really applicable to Wayfair. It’s one thing to answer a search query for a customer searching for a six pack of AAA batteries. In this case, customers can easily describe what they are looking for. However describing a product accurately is not so simple for customers looking to buy a piece of furniture for their home.
Imagine you wanted to buy the perfect accent chair for your space. How would you describe your ideal size? What words would you use to describe the perfect style? How would you describe the material or the comfort level of the cushion?
Furniture items possess more dimensionalities than items in a catalog for fast moving consumer goods. We need to leverage science to understand the products we have in our catalog, and also to understand the right products to surface to customers so that they can create their unique sense of home.
Being able to surface the right items to customers also has downstream impacts for Wayfair. Similar to when they are buying items of clothing, customers want to touch and feel items. This is something they are not able to do online. As a result, they often return items after receiving them in the mail.
However, returning clothing is far cheaper than returning a heavy item of furniture. It’s absolutely critical for a company like Wayfair that we are able to recommend the right products to customers. Given the large size of most of our items, a high return rate would result in high shipping costs, and have a negative impact on our business.
This is just one example of how science is helping Wayfair solve everyday problems. We are leveraging the latest in the state of the art machine learning to help drive supplier business growth, reduce our shipping footprint, make supply chain optimization decisions and so much more – it wouldn’t be an overstatement to say that science is an integral part of everything that we do.
Q. Why should scientists be excited to work at Wayfair?
People are used to thinking of Wayfair as a home goods online retailer. But we’re working hard on enhancing that perception, and highlighting that it is an exciting place for scientists to do really meaningful work.
Our scientists have had publications and talks accepted at leading industry conferences like RecSys, ODSC and INFORMS. There’s usually a huge amount of interest sparked by our participation at these events, when scientists see the wide range of practical problems that we are solving with science, and the real-world impact that scientists are able to have at the company.
We’ve always used machine learning to determine how to best use our marketing dollars to acquire and retain customers in the most effective way. That’s a problem we’ve spent a lot of time on, and it’s one that’s critical to our space. However, as we’ve experienced tremendous growth in recent years, we’ve also leveraged advances in natural language understanding, computer vision and other areas to help match customers to the products they love – and to do this at scale.
Q. What are some scientific innovations that you find exciting?
Like I said before, I’ve always been driven not by innovation for innovation’s sake – but rather by how it can be used to have an impact on the lives of millions of our customers from around the world. We are doing some incredibly interesting things at Wayfair. For example, we have a team that’s working on ways to allow customers to better get a tactical feel of the objects they are browsing on our web and mobile properties.
We are also thinking about our play in the Metaverse, specifically, how do we make it really easy for customers to have a digital copy of the home in the cloud. This will allow you to experiment with the look,feel and placement of furniture in a virtual environment before you decide to buy it for your home. This ability to visualize furniture accurately can take away a lot of anxiety around the furniture buying process.
There are many other projects at the company that I am really excited about. They are all unified by a common and very meaningful mission – to help people create spaces that reflect who they are, what they need, and what they value, so that they can feel right at home.