Fashion is one the fastest growing and evolving industries, with new trends and consumer behaviour constantly changing. Because of this, more and more are turning to data science and analytics for help. Yes, even high brow brands (although some might not admit by just how much!). The primary way analytics is used to shape the fashion industry is the increasing use of data to help predict the direction of future trends. By doing this, brands are able to create not only what consumers want quickly, but whilst maintaining incredibly profitable margins.
But first, it’s important for fashion decision makers to understand their consumers. Luckily, e-commerce and social media has made it easier to gain rich consumer insights.
Gaining Consumer Insights
The online space has become a treasure trove of customer information with online shopping growing 3 times as fast as its brick-and-mortar counterparts and consumers readily sharing their likes, dislikes, recent purchases, etc. on social media whether done consciously or unconsciously.
To manage all this data, artificial intelligence is used to get consumer information in real-time. It monitors which items are being viewed, purchased, returned, and reviewed. AI also analyzes trends and conversations that go around social media to inform product decision makers.
These analytics are so efficient that brands, such as Fashion Nova, can go from design prototype to online post within a mere 24 hours! This popular e-commerce brand heavily relies on data from their social media engagement rate, such as likes and comments, when creating their apparel. They have no physical presence and conduct all their retail 100% exclusively online. This a major game changer in the industry and increases margins not only by reducing cost of retail space but also manage demand more flexibly online to avoid overstocking and tying up cash in inventory – we’ll talk about this again later.
Unlike Fashion Nova, Zara has many brick-and-mortar branches though also has a fast-fashion approach. Still, they have partnered with Jetlore, an AI-powered platform that analyzes the retailers products and its consumers behaviours and preferences to inform marketing and product creation. Zara also tries to get into the minds of consumers using their search box feature by implementing bidirectional search engine, a deep learning form of AI that understands how people really talk by learning from the context of the entire search query.
With regular data collection and observation, trends can be anticipated before they start showing up on the runway. Knowing this, brands can create or curate things that people like. This is where predictive technology comes in handy. Using what may seem like a bunch of unrelated information, AI is able to identify hidden correlations between different variables, such as how weather, colors, and sizes affects fashion trends and overall sales.
Knowing the right styles to release is one thing, but producing the right amount is just as important. Inventory that doesn’t sell is put on sale, causing less-profitable margins, or worse, left in the stock-room. Similarly, products that sell out too quickly, causing unsatisfied customers and loss of potential sales.
A great example for this is Boohoo. As an online only store, they don’t need to fill up shelves and racks. Instead, they produce “test batches” of styles in small quantities. Using analytics, they’re able to quickly learn and adapt to what their consumers like and produce stock that is close to the demand.
How do brands connect consumers to their products?
For brands to get their clothes from warehouse to closet, many fashion e-commerce sites have machine learning built into them. This is used to recommend products and alternatives to customers. The more a customer uses the website the better the intelligence gets at recommending products. This creates a more personalized shopping experience.
To further enhance this experience, brands use VR and AR to allow customers to try clothes on virtually. One app called DressingRoom by The Gap, lets customers create avatars that have custom measurements for trying on clothes.
Digitization of fashion
In the time of the Coronavirus, we can expect a lot more usage of VR and AR in fashion using 3D design, CGI, body mapping, and other technologies.
ASOS has launched “See My Fit” which uses Zeekit’s deep-learning algorithms and AI technology to virtually but realistically dress models. This eliminates the need for face-to-face photo shoots and helps customers make informed decisions.
For Burberry, they launched a new exciting and personalized luxury shopping experience by using an AR shopping tool through Google that replicates the in-store experience wherever the customer is! When a customer makes a Google search of a product, they can see its’ AR version in real size wherever they are.
These are just some applications of data science in the fashion industry which we can expect to change drastically in the near future as world events such as the COVID-19 pandemic accelerate digitization and data driven insight, even in a highly creative industry like fashion.
With richer data and more sophisticated technology we can expect a change in AI learning algorithms that give more inclusive and personalized experiences to consumers. This however, cannot be done by machine learning alone. To get closer to reality, machine learning models need to be fed more diverse datasets by those in the fashion and science industries for it to be useful for all kinds of people and industries.
Furthermore, with all the power of data science it still cannot create trends and new ideas. In fashion, it is not out of the ordinary for new and original styles to emerge and for people to divert to different styles quickly.
While human choice is highly affected and optimized by data science and analytics for prediction technology, humans ultimately put the “art” into that process and for now, ultimately have the final say.