Changing how we shop
Fashion retail is ripe for innovation as the traditional shopping experience has not evolved much in the last century. While online shopping has resolved some retail therapy woes (goodbye long lines, mean salespeople and overcrowded racks), the proliferation of apparel and accessories brands online has left shoppers with too many options and too much stress. A simple search for “going out dresses” in my size on asos.com returns 1102 styles. Overwhelmed? We all are, but we can’t afford a personal stylist to do the work for us.
Enter Stitch Fix. A personalized “shopper” that curates the styles you’ll love and delivers them to your door every month for only $20. Customers simply buy the items they like and return the rest.
Stitch Fix has finally cracked how to customize the shopping experience at scale. The business model maintains an edge due to its distinctive operating model, without which the service would be defunct.
What’s in the business model?
The core value proposition is that each style is hand-picked for the customer by a fashion expert. When a customer signs up for the service she is paired with a stylist who receives information on her preferences and then picks the items for each Fix. Every Fix is therefore personalized, arriving with a note from the stylist in addition to fashion tips and tricks.
This service is not for savvy fashionistas – it is for women who don’t know what to wear and don’t have the time to figure it out. Stitch Fix creates value by teaching women how to dress themselves on trend. Every Fix arrives with style cards on how to wear each item, the company blog is laden with fashion content, and the website features an “Ask A Stylist” form where any woman can ask fashion-related questions and hear back from a stylist within a couple of days…all for free.
Stitch Fix also strives to make the experience as easy and seamless as possible. The styles are pushed directly to your doorstep, arriving like clockwork every month, requiring minimal effort from the customer. Each Fix includes exactly 5 items to reduce complexity in choices and decision-making, and it comes with a prepaid mailing envelope for easy returns.
Importantly, Stitch Fix has found a few ways to capture this value they’re creating. First, it leverages a subscription-based revenue model by charging $20/month for the service. This fee is offset when you keep an item, thereby enticing most customers to buy atleast one piece. It also offers a 25% discount if you keep all the items in the box, encouraging the undecided customer to just keep it all when she is unsure about an item or two.
What’s in the operating model?
Good Data, Better Algorithm
In reality, Stitch Fix stylists only do a fraction of the personalization work. Most styles are first recommended to the stylist by a proprietary algorithm, which makes targeted fashion recommendations. Stitch Fix uses a number of data points from each customer, including their feedback on each style to optimize predictions on a customer’s likelihood to purchase a given item.
When users sign up they share a host of information through a very detailed user profile survey. As one new user describes it on her blog: “The questions involved everything from my height and weight to how often I go out on dates, which jewelry tones I like, and how I prefer my clothes to fit each part of my body. [I was asked to] rate various styles, select my favorite and least favorite colors and patterns. At the end of the survey, I was asked to share a Pinterest board of styles I like, as well as links to my social media profiles.”
Furthermore every month, customers are asked to give feedback on all the pieces in their Fix. The website asks users to explain why they don’t like something. They are encouraged to use if statements (e.g., “this sweater would work if it was red”) to capture as much information about their tastes as possible. The algorithm then uses natural language processing to understand this written feedback and refine its recommendations. For example, when a woman writes that she is in her “third trimester”, the algorithm determines it needs to recommend the maternity line.
This use of predictive technology based on rigorous data collection thus reduces the burden of personalization on the individual stylist, whose time is costly and who cannot process as much information at scale as the algorithm can. Instead the stylist uses the algorithm’s recommendations and adds a “human touch” to the selection based on her discretion and knowledge of the customers needs.
“The algorithm will often be better than a stylist in determining what a customer will like.”
– Katrina Lake, CEO, Stitch Fix (HBS 2011)
Calling All Stylists
The company also has a unique HR strategy to attract as many stylists as possible. Stylists are allowed to work flexible hours from remote locations. In fact a large number of the 1000+ stylists on the company’s payrolls primarily work from home. This allows Stitch Fix to access a talent pool that may not otherwise be in the workforce, and it continues to expand its supply of stylists as demand for the service grows.
Poaching Top Talent
Stitch Fix has also invested in top talent in data science, e-commerce operations, and digital marketing by poaching leaders in these fields to join their nascent start-up. On the executive team you will find: Eric Colson who helped define the recommendation engine at Netflix, Mike Smith, the former COO of Walmart.com, and Julie Bornstein, the former Head of Marketing and Digital initiatives at Sephora. Stitch Fix has therefore positioned itself for strong growth by building an executive team with deep expertise in core areas of the business.
Why does it work well together?
Stitch Fix is a company where the process of value creation and capture relies entirely on the execution of its operating model. That is, customers will only subscribe to the service if the algorithm + stylist together make suggestions that the customer likes and buys. Therefore it is essential that the company deploy its resources as effectively as possible. This involves involves collecting as much data as possible on every user in order to make better predictions, and then using the limited resource (i.e., the stylist) to fill gaps where the technology is ineffective.
The human resource is positioned as the key service provider, but in actuality it is used sparingly to enable the business to run cost-efficiently at scale. Customers submit detailed feedback on every Fix because they think they are interacting with a stylist, sharing details on weight loss journeys and self image aspirations, all of which are captured by the algorithm. Every time a customer buys something or sends a note back to the stylist, the algorithm gets better and better at making suggestions for all its customers.
As a result, the business maintains an edge because it is in a cycle of continuous improvement. Every time it creates or capture some value, the operating model improves its ability to capture additional value in the future.