How Stitch Fix Uses Data Science and Machine Learning to Deliver Personalization at Scale

brought to you by WBR Insights

Online personal styling subscription service Stitch Fix is disrupting the fashion retail industry in a big way. Delivering highly personalized clothing recommendations right to customers' front doors on a regular basis, Stitch Fix eliminates the need for its subscribers to go out and shop - or even browse online - for clothing, and it's proving to be a big hit with millions of consumers. As CEO Katrina Lake puts it: "We send you clothing and accessories we think you'll like; you keep the items you want and send the others back."

From the customer's perspective, it really is as simple as that. Under the hood, however, there's a lot more to it, and the secret sauce is data science and algorithms.

Leveraging Data to Deliver Personalization

Stitch Fix employs a team of dozens of data scientists to deliver personalization on a massive scale. Machine learning algorithms are at the core of the company's business model, underpinning everything from client styling and logistics to inventory management and product design.

But of course, no matter how many magic formulas and algorithms are at work, they will only ever be as good as the data that drives them - and data is where Stitch Fix really excels.

The first dataset comes from Stich Fix customers themselves, who fill out an in-depth profile when signing up for the service. Questions range from fundamentals (height, size, and weight), to taste and preferences (do you like your shirts tucked or untucked?), to personal traits (are you a risk taker?), and lifestyle (are you a new mom?).

Maximize Your Visibility on Google with Our SEM/SEO Guide,
"Embracing the Journey to Discovery: The New Rules of Retail Search & Acquisition"

The second set of data is about the merchandise itself. Each item of clothing in Stitch Fix's inventory is tagged multiple times with "match scores" derived from client preferences, and then ranked. Even more data is collected from customer feedback, which is solicited every time a customer receives a curated box - known as a "Fix" - in order to glean insights as to how each item matched their style, fit, and price preferences.

Having all this data means Stitch Fix's machine learning algorithms can match the right products with the right clients. The algorithms figure out how much the characteristics of particular items - color, material, style, etc. - matter to each customer and a fashion box is curated from the results.

"Data science isn't woven into our culture; it is our culture," said Lake. "We started with it at the heart of the business, rather than adding it to a traditional organizational structure, and built the company's algorithms around our clients and their needs. We employ more than 80 data scientists, the majority of whom have PhDs in quantitative fields such as math, neuroscience, statistics, and astrophysics. Data science reports directly to me, and Stitch Fix wouldn't exist without data science. It's that simple."

A Unique Combination of Data Science and Human Judgement

While data is undeniably crucial to Stitch Fix, so are human beings. While machines and algorithms provide the initial filters, ultimately, the human stylists that Stitch Fix employs have the power to override the product assortment the styling algorithm has delivered and are key to understanding the nuances of customer requests and ensuring their experiences are personalized.

For example, if a client was to make a very specific request - such as "I need a dress for an outdoor wedding in July" - Stitch Fix's stylists immediately know what options would work for that event. Similarly, if a customer shares some intimate details of a pregnancy, a major weight loss, or new job opportunity, it's human stylists that know just how special and crucial life moments like these are - more so than even the most sophisticated machine - and will go above and beyond to curate the right look, connect with the customer, and improvise as appropriate.

"At the core of what we do is a unique combination of data science and human judgement," said Chief Technical Officer at Stitch Fix, Cathy Polinsky, "Our human stylists make our algorithms better and our machine learning helps our stylists perform better. By combining the art and science of styling we're able to create a far better client experience than anyone else in retail."

Final Thoughts

Stitch Fix was founded in 2011 and went public in November 2017. The company sold $730 million worth of clothing in 2016, $977 million in 2017, and forecasts that it will sell $1.23 billion in 2018. To spur growth, it is introducing an on-demand styling service for kids, which follows expansions into men's and plus-sized fashion categories last year.

Using its special blend of data, machine learning, and the human touch, Lake is confident that Stitch Fix will maintain its competitive edge and earn even greater brand loyalty the better both its stylists and algorithms become at determining the preferred style for each and every customer served.

"It's simple: A good person plus a good algorithm is far superior to the best person or the best algorithm alone," said Lake. "We aren't pitting people and data against each other. We need them to work together. We're not training machines to behave like humans, and we're certainly not training humans to behave like machines. And we all need to acknowledge that we're fallible - the stylist, the data scientist, me. We're all wrong sometimes - even the algorithm. The important thing is that we keep learning from that."

You can hear Mike Duboe, Head of Growth at Stitch Fix, speak at eTail West 2019, taking place next February at the JW Marriott, Palm Springs, CA.

Download the Agenda for more information and insights.

Return to Blog