So why have apparel brands lagged in doing the same? In many cases, identifying the “next big thing” — what will sell, and the rate at which it will fly off the shelves — is still steeped in guesswork and unsupported instinct.
For brands, the use case is abundantly clear. The global apparel market is one of the world's biggest sectors. It's valued at $3 trillion and accounts for 2 percent of the world's gross domestic product. This industry is trying to find its footing in a landscape of constant change, driven by new technologies and consumer spending that's moving away from brick-and-mortar to online.
In the Unites States alone, there were more than 211 million digital shoppers in 2016 who browsed through mountains of information, from new products, pricing shifts, promotions and more. This data, when amalgamated, could be used to provide retailers with a new, in-depth way of exploring the opportunities hidden within the retail landscape.
That sort of on-demand knowledge could show you which dresses priced between $10 to $30 sold best last week — and could be the difference between blindly buying into a declining trend and avoiding it altogether. The possibilities are numerous. But even if you have access to this data, the question is: how can you take advantage of it to make critical decisions, increase customer loyalty and boost sales?
Big data, big opportunity
Big data is exactly what it sounds like: information on a large scale. But more commonly it means a large collection of structured or unstructured data that is pieced together by computers and organized in a way that makes it possible for humans to derive valuable insights rapidly. Used in this way, big data can be the key used to answer previously unanswerable retail questions about what people are buying, the prices they're paying for them, and when are they buying.
As such, retailers can make better pricing and assortment decisions, reduce markdowns and decrease costs of dead stock by analyzing what's happening in real time or over a specific period. The focus of the analysis can be as broad as the entire world or narrowed to a single category, sub-category or trend.
At EDITED, we launched a filter in October last year that enabled retailers to drill into the rapidly growing athleisure and activewear space. Using the tool, companies can compare activewear versus non-active wear products, review specific keywords to understand micro trends and opportunities, and spot top-moving items to take immediate action on promotions, pricing and replenishing. In a category predicted to skyrocket in the next decade, this tool can give retailers an essential edge that pays off big time.
Moving to machine learning
Consumer tastes are highly changeable and brands face greater competition in an increasingly saturated market, so big data alone is not enough to make a tangible difference. To compete effectively and grow, retailers need to rely on their own expertise as much as the data they see before them. They need to be able to ask the right questions and know how to follow the answers to see where they lead. Data may be able to show retailers more than they'd ever imagined, but there's no autopilot.
To obtain that data, machine learning — and recent advancements in artificial intelligence (AI) — can provide retailers with powerful tools to maximize the use of data to make such an impact. These models can use supervised learning to distinguish one product type from another using a curated test set. Models such as neural networks use supervised learning techniques to teach themselves to derive patterns and traits in complicated data by mimicking how our brains process information. Such techniques can be used to recognize specific patterns and shapes when processing apparel images.
In fact, some of the most powerful machine learning algorithms today are applied to areas such as extracting color from product images, analyzing ambiguous children's wear sizing across a myriad of brands, or determining whether a pair of tights belongs to a sportswear or lifestyle category. The point of each is to give retailers access to the data, without browsing competitors' sites or sneaking into their stores and hastily trying to count items. With this sort of data, they can jump right in and do what they're good at: retailing.
AI today, and tomorrow
Developments in computer science, mathematics and statistics are often industry neutral, and the retail sector is just one of many others that are taking advantage of major technology advancements. Driven by the massive growth in online shopping, it is now possible to analyze consumer demand as well as retail assortments, pricing and approaches.
But what's next? In recent years, developments in AI have opened the floodgates to many more opportunities. At EDITED we've recently been caught up in the magic of neural networks, models based on the behavior and structure of neurons in our brains. These technologies can teach themselves to derive patterns and traits in complicated data by mimicking how our brains process information. Who knows? One day they could be applied to detect patterns in fashion images, giving new meaning to instant runway trends.
Even more recently the machine learning community has been excited about Generative Adversarial Networks. These models combine the idea of neural networks with game theory, resulting in powerful problem-solving models. Machine learning is a rapidly advancing area and each development leads us further along the AI spectrum towards systems that have even greater capabilities in learning, understanding and reasoning. The real-world applications for retail are yet to be seen, but the possibilities are boundless.
AI in retail is already being applied to provide insights earlier generations of retailers never would have dreamt of. From recognizing patterns to classifying sportswear, AI helps its users generate data-driven insights across terabytes of data. But this is just the beginning. In the future, your wardrobe might make an outfit suggestion based on the weather. You might be able to buy clothing labels designed by computers. You might be told by your computer if that shirt doesn't go with those shoes.
How this will define the future of competition in retail remains to be seen, but one thing is for sure: the days of the quarterly comp shop are gone. Long gone.
Sophie Coy is a retail data manager, and Joe Berry is a data scientist at EDITED, a retail analytics company.