Ask a Data Scientist at Rent the Runway
Anna Smith works as a data scientist for Rent the Runway, after filling a similar role at Bitly. With such a strong data background to her credit, you’d think that she’s had years of training and planning to get to where she is today, but that’s only part of the truth.
Anna actually got her start at Bitly when she accepted a summer internship at the company as a graduate student in physics. She loved the work so much that she signed on as a full-time employee and never looked back.
I recently sat down with Anna to discuss her background, her work at Bitly and her current role as a data scientist for Rent the Runway.
Moving from physics research to a full-time data science position may look like a bit of a stretch, but as Anna explains it, the principles behind the two are the same. She started graduate school wanting to study quantum computers.
After a while, however, she realized she was more interested in the theory aspect of it, and so she worked with the computer science department to bring both computer science and physics together in order to address various problems.
Anna accepted the Bitly interview during her graduate studies out of curiosity. After arriving, she realized that the hands-on nature of the company allowed her to learn faster than ever before. Since she was working with datasets on a daily basis, she picked up skills very quickly, and realized that it was a process she enjoyed — maybe even more than she enjoyed her physics studies.
Bitly was the start of Anna’s transition from academic physics to corporate data science. If you share links across the web, you’ve probably used Bitly before. This URL shortening service is used to save, share and discover links. It has long been popular on Twitter due to its abbreviated nature. Bitly gained even more traction when it became Twitter’s default URL shortening service back in May 2009.
As a marketer or community manager, you might use Bitly to create vanity URLs while tracking certain campaigns. Digital strategists might use Bitly to save bundles of links and share and track them on social media. More recently, Bitly has even expanded to offer an analytics platform that helps web publishers and brands grow their social media traffic.
At Bitly, Anna was passionate about telling the story of the data she interpreted from Bitly’s shared URLs. Part of the appeal of learning in a fast-paced environment like Bitly was the trial and error process, and being able to interact with others — including four or five other data scientists who often served as mentors, Anna says.
The data team at Bitly tries to make sense of how people use the Internet by looking at data that customers create. How is this data generated? Information is compiled every time someone shortens a URL with Bitly or clicks on Bitly links. The information can then be analyzed for larger patterns or trends.
From these datasets, Bitly engineers can create a product that customers might not even realize they need, or simply share their findings on their blog for others to learn from. They might even connect with business partners to see if their findings would be useful to others, even if Bitly cannot use it at the current time.
Though Anna enjoyed her graduate work, she was used to working in groups of just one or two, where she could make her way through an entire project with only her advisor to bounce ideas off of. Graduate school also involved a lot of “working in a dark room,” as she puts it, waiting for results. At Bitly, where Anna worked with data science, she was able to see her work become usable really quickly, which was especially rewarding.
Rent the Runway
Now, Anna works at Rent the Runway, an online company that provides rentals of designer dresses to those who might otherwise not be able to afford them or may not need them long-term. The Rent the Runway model is simple, and customers are loyal.
How does it work? When they know they have an upcoming formal event like a wedding, bachelorette, gala, military ball or New Year’s party, customers simply place an order for a dress of their choosing and the company mails it straight to them. Rent the Runway provides free shipping, handling and cleaning for all dresses. The customer then has four days to wear and return the dress, and is ultimately charged a nominal fee for the rental, especially considering the original price of these high-end garments.
So where does a data scientist fit into an online fashion company? It turns out that Anna Smith has plenty of work to keep her busy. She currently works with multiple teams, focusing on:
- Pure analytics: Which pages see the most interactions and which are least likely to convert? Are customers using the dress reservation calendar? What size or style of dresses do they look at?
- Recommendation system: Are customers clicking through/investing in other dresses or accessories that are dynamically recommended to them, or does the algorithm need to be further refined?
- Site experiments: Which site design converts best? How well does this style of dress, discount or other variable perform in A/B testing?
- Machine learning problems: Which customers are most likely not to return a dress? How can this be prevented?
As for specializing in a particular area, Anna says that her work is all over the map right now. Since she’s new to the company, her analysis is largely exploratory, and she helps Rent the Runway understand what the data has to say. She tries to answer hypothetical questions: Should Rent the Runway scientists assess data from anonymous (not registered) users? Are the page suggestions of what a customer has looked or ordered before helpful, or just a waste of space? Should the site stop displaying a dress to a particular customer once they’ve already seen it? All of these answers can inform website redesign, customer service and even the company’s investment in its own inventory.
In her career, Anna Smith has gone from graduate student in physics to working data for a link bundling/online company to analyzing data for a fashion-on-call website. However, as she’ll tell you, the key element that tied all of these jobs together was an innate sense of curiosity, intelligence, and social skills. She firmly believes that a data scientist is not an engineer in search of a direct path to get from A to B, but rather someone who likes to look at problems and create solutions, however tangential they might be. Above all else, a data scientist is simply curious.