Stitch Fix: Personalizing Fashion with Data
As we have seen in lately reports, digital transformation is disrupting the retail industry. E-commerce sites are changing the consumer’s behavior, shifting purchases from brick and mortar stores to on-line. In the apparel industry, Stitch Fix created a business model that uses data to offer, not only clothes, but personal recommendations in style. Are we in front of a new winner?
How does Stitch Fix work?
Stitch Fix is an on-line clothing shop that provides personal styling services at affordable costs . When signing up, customers need to fill in their personal profile and respond to a small questionnaire of their fashion preferences: Does the customer like formal or casual style? What are her favorite colors? Is she willing to take risks? Also, customers are required to share price preferences, body type, weight, bra size, and their Pinterest account. The company uses customer’s data, to create algorithms that feed the recommendation engine. Artificial intelligence is used to create an initial list of recommendations, and then a human stylist will reduce the list to 5 pieces that will be shipped to the customer.
Customers need to pay $20 per fix, but this amount will be credited if at least one item is purchased. Customers can try the pieces at home, keep the things that they like, and return the ones they don’t. Stitch Fix encourages customers to provide feedback on each item they received. This serves as additional information to refine the algorithm and improve client’s satisfaction. 
With this model, Stitch Fix is creating value by:
- Reducing shopping time. “A few years ago, market research firm OnePoll surveyed 2,000 women and found that each year women spend more than 100 hours on 30 trips to shop for clothes, 15 shoe-shopping excursions taking 40 hours, and a full 50 hours per year window shopping.”  By eliminating the need of buying, customers can spend their time in other activities such as working or going to the gym.
- Making it easier for customers to follow latest fashion trends. For some people it is super easy to be on top of the latest trends, and love spending time following bloggers and Pinterest boards. But for others, it is hard to keep up with them. Stitch Fix solves this problem by advising customers based on their preferences and latest trends.
- Increasing overall shopping expenses. All of us have at least one friend who hates shopping clothes. Some people don’t like going to stores, and hate trying a lot of styles and not finding what they want. By “pushing” products to customers, Stitch Fix is increasing consumption in this specific target segment, increasing the overall size of the apparel market.
Stitch Fix has more than 2.2 million active clients, has sales nearly at $1B and raised $120M in its IPO [4,5]. Even though it seems to be a winner in the digital era, there are still some challenges that the company needs to solve. For example, can artificial intelligence be enough to recommend outfits to customers? How many customers are willing to adapt this new purchasing habit? How scalable is the business model? Are the any opportunities to partner with brands and have new sources of revenues?
Student comments on Stitch Fix: Personalizing Fashion with Data
Great post – thanks D! Definitely think the Stitch Fix’s business model is an interesting one and they do seem to be creating a lot of value if the issue for consumers really is time savings. But a few questions related to consumer behavior that I think Stitch Fix needs to better address before it can be labeled as a ‘winner’: how large is the market for Stitch Fix? Will most consumers actually be willing to change their shopping habits to have someone else do their shopping? Will consumers be willing to give up the experience of shopping with their friends or shopping for stress relief? Are consumers taking away part of curating their own identities by allowing others to essentially curate their style over time for them? On this last point, so many Instagram bloggers have found success in helping women learn and explore their styles.
Lovely for time-sensitive customers and basics-fans like myself, but how do we prevent the ultimate commoditization of fashion? What are relevant inputs that will allow AIs — or designers? — to find that ‘perfect fit’? Isn’t there a special bond with that one piece that you found in the middle of that rag, and that now you can’t live without?
While I think many retail and fashion search engines are putting lots of resources in finding the perfect outfit for YOU, by basing that recommendation on other (past) purchases/searches (ejem… I HATE AMAZON’S RECOMMENDATIONS — when will they understand that if I bought a pair of Nikes 2 months ago, I won’t be needing another pair any time soon!), I would suggest they should be thinking more outside the box!
The question they need to answer is when and what you need to buy! What is the root cause of our impulse purchases?
Wouldn’t it be great if those AIs could uderstand some of the following variables?:
– Occasions: When is our birthday, or when did we broke up a relationship, first day of a new job, or a promotion, the date for our salary deposit, our best friends weddings, our anniversary
– WTP or Budget: This can usually be an input or derived from other purchases… but what if AI knew what credit cards you hold, which discounts apply to each of them, where you accumulate miles, when its paying cycle closes, and optimized for all that!
– Role models / Influencers: Who do we like, look up to, wanna imitate? Which famous do we follow on IG? Which firend would we take as a shopper advisor? Which are the brands we enjoy, the places we frequent, where do we do our hair? What musician or sports person drives you crazy?
**Spoiler alert — For all that info JUST CONNECT TO ACCOUNT TO YOUR IG/FB!**
I believe there is a huge opportunity in understanding psichology and behavioral economics for fashion recommendation algorithims! Come on AI, come appeal to my emotions!