Artificial Intelligence – the Future of the Online Fashion Industry

Artificial Intelligence is really not as scary as it sounds. A far cry from robots putting millions of hard-working Americans out of work, AI is becoming a valuable breakthrough for both retailers and customers. Underneath the broad category of AI lie four major applications that bring incredible value to the world of digital commerce: voice search, hyper-targeting, personalization, and chatbots.

giphy20

Voice search became mainstream with Apple’s Siri, but Amazon has utilized it for increased sales through Echo. Through these AI assistant devices, neural networks take in auditory data and attempt to respond in the same way a human would. While these are primarily transactional in their current functionality, the Four Horsemen and many of the largest retailers are heavily investing to expand the possibilities of “AI Assistants”. Hyper-targeting brings value to retailers by collecting and analyzing tons of data through machine learning, a subcategory of AI. Our phones can now point out a store with umbrellas in stock that we will pass on our way home from work when the forecast calls for rain. (Okay, I do actually find this one a bit scary.)

AI enhances customer experience through personalization and chatbots. It enables faster and more precise consumer recommendations using more than just your purchase history. In this way, AI combines data from purchase trends, customer loyalty, demographics, and browsing patterns in order to give customers exactly what they want exactly when they want it. Chatbots are getting better at understanding a wide variety of questions and comments, no longer sending pre-programmed responses to simple questions. Overall, AI software is expected to bring revenue of nearly $60 billion to online retail by 2025.

One sector of retail that is heavily investing in artificial intelligence is the online fashion industry. If you’ve never heard of Stitch Fix, it is an online subscription and personal shopping service. Basically, you set up a profile on the website or app which asks questions to determine your fashion style, clothing sizes, and price preferences. Next you request a “fix” whenever you want and a box with 5 items will be shipped to your door. After trying them on, you can keep what you like and send the rest back. You only pay for the items you keep. There is a $20 styling fee per box but you can apply that amount toward the items you end up purchasing. It’s a simple idea that solved a previously unmet need – the ability to try new clothes with the convenience of online shopping without the commitment of purchasing before trying. Many people, myself included, will order something online that doesn’t quite meet our expectations upon delivery, but the hassle of shipping it back for a return or exchange does not seem worth the effort. This results in dedicating a “I may wear this one day” corner of the closet until it grows too large and we decide it’s time for some spring cleaning. Eventually all of those not-quite-right, never-been-worn, tag-still-attached items will be folded up nicely and placed in a shopping bag for their trip to the Salvation Army donation box.

Stitch Fix fixes all of that! As impressive as this business idea is, their execution is the really remarkable piece. Stitch Fix uses a seamless combination of artificial intelligence and human expertise for their business operations. Specifically, it uses machine learning and natural language processing together with the massive amounts of data compiled from the style profiles and purchase/return history of each customer. After the Stitch Fix algorithm analyzes the profile results, personal notes and more nuanced data, it sends the results to a real-life human person (aka “stylists”). The stylist then makes the selection of 5 items for the fix box, packages them with a personal message and sends it out. While the technology helps increase productivity and efficiency, human interaction throughout the process is crucial. This combination helps to more accurately gain insight to the customer and maintain stronger relationships with the brand.

Just in case Stitch Fix’s $ 1.22 billion expected 2018 revenue isn’t enough to convince you that artificial intelligence is the future of fashion, Amazon is on board now too. Engineers at Amazon are working on algorithms for an AI fashion designer. As they aim to increase their footprint in the fashion industry (they even recently started their own line of apparel) Amazon hopes their technology will even be able to predict fashion trends of the future. Using a GAN (generative adversarial network) they can analyze properties of particular styles by looking at many images and apply the styles to an existing item. Leave it to Amazon to take AI to the next level!  While artificial intelligence clearly has a prominent place in the future of e-commerce, there should still be a place for the human experts as well.

12 comments

  1. It has been interesting to watch the fashion industry evolve over the last decade, and I’m sure the next will bring even more dramatic changes. The level of targeting made possible through data grows more shocking by the day, demonstrated clearly by your example of the phone and stores with umbrellas. It honestly can be a bit scary. I agree that Stitch Fix’s business model is quite impressive. While the idea of online shopping may not appeal to my mother and some of her peers, its use is ubiquitous across my generation. My friends and I all struggle with the “I may wear this one day” section of our closets but seem unable to shake the allure of online ordering. Stitch Fix recognized this challenge, saw a gap in the market, and went after it. I had not heard about Amazon’s ambitions in the fashion industry, but I am not surprised. While the Amazon name does not currently make me think of fashion, they seem to have all the tools needed to succeed.

  2. I think you are so right that they key to Stitch Fix’s success has been the combination of AI and personalization with human connection. I know a few people who used Stitch Fix at the beginning and found that they did not like a lot of the items sent to them. I think especially with clothing, our styles are constantly changing and, at least definitely for me, are not always rational or make sense. Having a human or ‘stylist’ that is interacting with the data and personalized information is vital because it allows them to notice the nuances in style people may have that data and AI may not be able to pick up. I am interested in where people see the future of AI going – will one day AI and algorithms be powerful enough to eliminate the need for humans or is human connection and interaction always going to be needed alongside the technology? I tend to lean towards the power and need for humans combined with technology, but technology is clearly becoming more powerful and precise at the same time.

  3. I’ve seen a lot of commercials recently for Stitch Fix but I would never have thought that AI is working behind the scenes of this service. I think the platform also shows the potential for network effects — as more people starting purchasing through Stitch Fix, I’d imagine that the company will be able to attract more brands. As for Amazon developing an AI fashion designer of their own, I agree with Molly and it will be interesting to see if they can break into the fashion industry.

  4. I haven’t heard of Stitch Fix before but your post makes me want to try it out! However, Katie’s point of our changing style choices really resonates with me and I think it will continue to be crucial for a human to sort through my fashion choices in order to adapt my personalized box. I also think this company’s idea caters to some of us that get overwhelmed with making a decision of what to get and then end up getting nothing. I feel like as I get older, I have gotten pickier and pickier so having someone making that decision process smoother be narrowing the options would really help me feel more decisive. It makes the process much less overwhelming. Having to pick from a box of five items at a time is much easier than picking through a whole mall. It will be interesting to see how Amazon will carve into the fashion industry using AI. When I think Amazon, I definitely don’t think fashion so I’m looking forward to discovering how they compare with other companies whose names you can trust for fashion.

  5. Like Katie and Jenny, I also believe that the AI-human combination is the perfect implementation of automation in the fashion industry. Since this is such a subjective and aesthetic-heavy field, algorithms won’t (yet) be able to perfectly predict the wants of clients. By adding human input at the very end of the search process, Stitch Fix can drastically cut down on time spent searching while still delivering accurate and “will-wear” results for the customer.

    Since this is also a purely online company, I’m sure their algorithms will adapt over time as you remain a customer, as you have also stated. I can only see services that provide such convenience becoming increasingly popular over time, as customers spend less time crowding racks and aisles and more time doing more fun or productive activities. To me, all this means more trouble for the traditional retail outlet, which must radically reposition itself in order to stay relevant in an increasingly AI-dominated world. Current trends don’t give retailers much confidence, as many have shuttered their doors, but only time will tell.

  6. Very interesting post, I actually tried to stitch fix after reading your article, they ask me to do a quiz including Fit, Style, Size, which collects detailed body sized data and style preference to help me to find the best fashion option. The experience is fun and I might get some great outfits.
    For the Online fashion industry, I think there is a gap between the customer’s expectation and the products they received. The existence of the gap is because the buyer can’t get a close look at the products, and they can’t try them on, in addition, there is no customer service to help your selection, it’s all up to you. But Stitch Fix is using the AI technique to narrow this Gap, and I think such industry will have huge market potential.

  7. I’m extremely interested to see where else this technology will be implemented in the consumer market besides that of music and and clothing. Although I believe online shopping is going to be the shopping of the future I agree with @mollypighini when she said that she is doubtful that this style of service will appeal to older generations. It seems that the process going to a store and seeing all your potential options is something that people still really enjoy. I’m surprised other large corporations haven’t begun to utilize this technology and wonder who will be next. Great post!

  8. Nice post. We’ll be dealing with AI in a couple of weeks. It’s certainly poised to be the most disruptive technology of the next generation. Hadn’t considered the fashion implications before!

  9. That is a very interesting post. I have never heard about Stitch Fix before, but it seems it can make shopping more convenient and experimentative. After all it may offer you new style that you would never choose yourself. I wonder how many items shipped by Stitch Fix are actually kept by customers. I assume the service also gets better as AI collects more data about each consumer’s preferences. Personally, I always prefer to shop in store because I could never tell how an item will fit me. I wonder if one day AI will be able to solve my problem and just show me exactly what I should buy. It seems unlikely today, but given the rate of growth of AI it is probably coming faster than I can imagine.

  10. Personally, I know that L shirts always fit me no matter what brand they are, for example, and so I don’t have to go through all of the worry that I know a lot of people go through when trying to online shop. But I definitely see the huge upside to this market and industry. These AIs are able to process so much more data, so much more efficiently, than humans could ever dream of doing. Adding the human stylist at the end is a really nice touch, because there are some things that even the most advanced AIs could not do. The human connection is what people want out of this. AI is going to continue to adapt and be applied to every aspect of life, and these early movers are giving themselves huge advantages.

  11. I found this post really interesting and I have previously looked into Stitch Fix because I think it’s a really cool service. That said, the reason I’ve never actually used it is because I try to online shop only when I actually need something and I don’t really have the disposable income to shop for whatever someone thinks I might like, much as I may find some really interesting pieces that way. That said, I would be infinitely more attracted to a service that would ask me for all of that information and then make recommendations to me that I could choose from for myself. I know that I recently interacted with a similar service on a department store’s site (I think Nordstrom) and while I loved the idea, it didn’t particularly help the process. That said, I’m looking forward to watching how this technology evolves and improves in the future!

    1. Another great thing that Stitch Fix has evolved over time is the ability to request a “fix” whenever you want. It is no longer a pre-set subscription service. In addition you can set up your profile to request certain types of clothing items (pants, dresses, tops, etc.) and list your per-item budget. They have made it much more flexible for the users to dial in on receiving items they want, not just items the stylists think they might want.

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s

%d bloggers like this: