Style is subjective. One moment the newest dress style can be on every visible body in a hot nightclub and within a few months its doomed to sale racks and thrift shops until it comes back in style again. There's no exact way to predict what the next fashion trend may be, except for the uncertainty that most ideas come back in style after a while in one form or another. Betting on the next fashion trend is risky business, meaning the difference between a profitable quarter and empty racks or an overflow of unwanted inventory and no customers. Computers could be the fashion world's next great bet on what's the next up and coming trend.
Stitch Fix, a San Francisco based company, offers customers 5 personalized outfits picked based on their fashion preferences. The company uses a combination of data analytics as well as personalized stylists to curate a box of 5 items for their customers on a monthly basis for them to rent or purchase. Customers can try out the clothing, and if they like the items they can buy them at a discount of up to 25% off the entire box price if they buy all of the clothes for that month. Stitch Fix sources their clothes from lesser known, independent brands, with all clothing and accessories fitting within the price range of $28-$188. The computer uses analytics to pick out forms, sleeve types, prints, necklines, and clothing details that it believes will resonate the best with consumers. It's predictions have been so well that three new tops, designed partly by computer analytics, sold out in 2016.
The rise of subscription boxes has grown exponentially within the past two years, by 3000% to be exact, but the use of big data analytics within the fashion realm to predict new trends is relatively new. One method of data collecting and analytics is catwalk analytics, implemented by Francesca Muston, WGSN's head of retail and product analysis. Catwalk analysis requires noting data on what shapes and types of garments are hitting the runway as well as the colors, prints and details that are there as well. Just like any data analytics, it's about finding patters: what's the most popular item, shape, color, etc of the season and using this information to create smarter inventory and marketing decisions. A more technical form of analytics in the data mining field is cognitive computing which aim to mimic decision making, understanding, and comprehension of the human brain while also recognizing patterns within larger sets of data. This form of computing is especially marketable for the analysis of unstructured data, data not housed in databases, such as information gathered from mobile phones, social media, images, audio and YouTube.
Although we may not be scrolling through our Instagram's to view IBM's Watson modeling its newest picks for fashion trends, we'll be seeing the field of big data analytics merging more with forecasting in the fashion world. Computer's aren't fool-proof, so inevitably there will be some questionable fashion choices, but in that way computers will have thier opinions just like we have ours on what's an in-style piece.