Erika Frost says that science “was in the air” at her childhood home in Colorado.
“My parents were software engineers,” Frost says. “In fact, my mother went back to college to get her Ph.D when I was five years old. Growing up, I heard her talk about her work in areas like biomimicry and aeronautical research. It was all very inspiring.”
When she was in her junior year, Frost discovered the scientific area that triggered both her passion and curiosity.
“I had a natural interest in subjects like the social sciences, behavioral psychology and societal dynamics. Economics allowed me to approach these areas through a quantitative lens. I was drawn to the field because it relied so heavily on Math.”
Frost enrolled at the University of Nevada, Reno in the economics department double majoring in math and economics. Today, she is pursuing a doctorate in economics at the University of Wisconsin, Madison.
An academic focus on industrial organization
As she pursued her doctorate, Frost focused on the area of industrial organization. Industrial organization deals with strategic behavior of firms in the market, and encompasses areas such as regulatory policy, antitrust and competition. More specifically, economists apply the theory of price to these criteria by analyzing how the market forces of supply and demand influence the price of a product or service at a particular moment in time.
Frost’s research focuses on the impact of acquisitions of brands like Honest Tea by larger corporations like Coca Cola on consumer welfare. She studies the impact of availability and market structure as the primary drivers of changes.
During the second year of her Ph.D. program, Frost and her colleagues began research for an academic paper titled Scalable Demand and Markups. The paper explores the effect of price markups on demand across a wide spectrum of consumer packaged goods markets from 2006-2018.
The team uses logit nested models to estimate demand. Logit models estimate the probability of an event (e.g. did the consumer buy, or not buy, a product) based on a number of independent variables. Logit models are especially relevant to studying market dynamics of goods marketed to consumers where the products are differentiated, there are a wide number of competing products, and consumers typically chose one of the products.
Frost’s academic work would prove to be especially relevant for her future internship at Wayfair.
From academia to the industry
During the fourth year of her doctorate, Frost’s academic advisor forwarded her a job posting Andrea Guglielmo. Guglielmo worked as a Senior Manager and Economist at Wayfair, and had graduated from the economics department at the University of Wisconsin, Madison.
Frost was intrigued by the job posting.
“The job description was broad,” says Frost. “It said that interns would be able to work in areas like industrial organization, causal inference estimation and time series forecasting. Industrial organization was aligned with my research work, and I applied for the role.”
Frost says that the interview itself was pretty straightforward. She estimates that half of the interview time was devoted to assessing her depth of knowledge in the area of industrial organization. The other half delved into her research along with the paper she was working on with her colleagues.
“What I particularly liked about the process was that Wayfair assigned me my project only after I got the job, and only after developing a deeper understanding of my interests.”
Real world impact at Wayfair
Since her interview, Frost has completed two summer internships at Wayfair. During both stints, her work was tied to Wayfair’s demand simulator tool.
Wayfair uses the demand simulator tool to model the impact of actions that lead to major price changes, such as a change to the pricing algorithm, a change in the overall catalog price level, or a change in how costs are estimated.
“My work is focused on estimating the impact of price increases,” said Frost. “My project took into account the probabilities for discrete events such as when the consumer buys a product at Wayfair, but can also buy the same product or a substitute from one of Wayfair’s competitors.”
When she commenced her project, Frost used a simple nested logit model – where all of Wayfair’s goods were placed in one “nest” or group, while outside groups were placed in another group.
To understand the patterns of substitution as they would play out in the real world, Frost went one step beyond by developing a general nested logit model. In the next iteration of her model, Wayfair’s products were broken out into separate groups (E.g. couches, red couches, storage organizers, etc.). All outside products were placed in a different group. To imbue the model with even more complexity that was reflective of the real world, the nesting parameters were allowed to differ between groups.
At every step of the journey, Frost examines how closely the outputs of the demand simulator align with the results of real world experiments.
I love how Wayfair sets you up for success at every step of the internship program
Frost says that the opportunity to make a real-world impact is one of the biggest reasons to pursue an internship at Wayfair. She says that she is able to make a direct connection between her efforts and its effect on customer satisfaction, because she has complete ownership over her project.
“I have one-on-one meetings with my manager and a mentor every week,” Frost says. “I can ask questions and my colleagues help me work through any obstacles I might face. I love how Wayfair sets you up for success at every step of the internship program.”
