Crafting a Digital Fashion Vocabulary

by Christoph Lange - 4 Apr 2017

Companies involved in the business of fashion like Zalando have shifting priorities, and technology has become more important than ever. Technology at its heart is incredibly granular – highly detailed, and composed of many complex parts. This description can also be applied to fashion: Trends, seasons, fabrics, textures – the entire industry is astonishing in its scope and depth.

Contextualising the world of fashion to create the most relevant digital footprint is at the core of our work when it comes to fashion insights. How can we map the trajectories of trends and stay relevant? How can companies be technically innovative while being en vogue and design-led in terms of usability and relevance? The way we communicate this mindset is key.

Consumers often struggle to describe the different types of fashion items they’re looking for, and while this is a very promising area of research, we need to also ensure we’re the experts when it comes to terminology and language. Software can combine an understanding of what we know about consumers and what we understand about them, but it’s equally important to stay on trend to account for fashion’s fast paced nature. This means expanding your range beyond mere words, including images, styles, and behaviour.

Fashion recommendations and trend forecasting have a technical underbelly that smart companies are already taking advantage of. To do so, they’re bringing together the worlds of data science and fashion to collect and create a digital fashion lexicon. What is the difference between a chevron stripe and pinstripe? How can e-commerce players recognise whether a certain curve in the jeans trend is on an upward or downward trajectory? Zalando is putting together and updating physical and digital fashion glossaries to better support their engineering teams and bolster research in the realm of trend forecasting.

By bringing in industry veterans of the fashion world and pairing them with product, engineering, and data science professionals, innovations such as image recognition can be optimised by deep learning to deliver truly relevant content for consumers of fashion. For example, we’ve integrated a Fashion Librarian into our Fashion Insights Centre in Dublin, who has utilised her background in the fashion industry to clarify fashion terminology for entity recognition. Explaining the difference between 'style' and 'look' clarified the focus for our data scientists and guided their research in a precise direction. It enabled them to understand and develop a better grasp of trend forecasting and the trajectory of trends.

Understanding the fashion industry and its transient nature is integral to forecasting, where we must stay on top of what is happening on a daily basis. This is one of the most important roles that a Fashion Librarian plays for Zalando – their constant pulse check on the latest in fashion ensures we’re kept in the e-commerce game. Following on, the translation and contextualisation of this information educates our engineers, whose job it is to create the best tools and services for our customers and the greater e-commerce world.

Fashion and technology make a powerful partnership. Their relationship has created a much wider understanding of both industries from either side, opening up opportunities that weren’t even foreseeable five years ago. In the realm of fashion recommendations and trend forecasting, subtle differences can make or break your attempts to innovate, which is why we’re setting the groundwork via a well-crafted digital fashion vocabulary. By looking after the smallest of details, you’ll be making the most fashionable of impressions – you just need to know the lingo.

Similar blog posts