Is fashion big data compatible?
The latest round table of the Lectra-ESCP Europe Chair was held on October 7 in Paris, during Fashion Tech Week. This year, the theme was the opportunities offered by big data to fashion companies. How will they take advantage of it? Valérie Moatti and Céline Abecassis-Moedas, academic co-directors of the Chair, invited a range of panelists to debate the use of data in the fashion supply chain: François Nguyen, Chief Data Officer for La Redoute, Nicolas Santi-Weil, CEO of Parisian men's brand AMI, Edouard Fonkenell, founder of marketing agency Claravista, Benoît Sabatier, CEO and co-founder of the startup Clother, and Maximilien Abadie, Strategy Director for Lectra.
l-r: Valérie Moatti, Maximilien Abadie, François Nguyen, Edouard Fonkenell, Nicolas Santi-Weil, Benoît Sabatier
Is big data compatible with the world of fashion?
An emotional concept, fashion is the realm of the imagination, the ephemeral. Data, on the other hand, is strictly rational. Although these two worlds seem to be quite far apart, they are not incompatible, explains Maximilien Abadie. He is convinced that big data can help fashion companies without harming their creativity. “We are passionate about the product,” agrees François Nguyen. “Data can push designers even further in their vision of the product".
What are the benefits of big data, in the fashion industry?
There are three lines of action for big data, says Edouard Fonkenell. The first is to improve customer insight to understand what they do, what they like and what they think of us. Out of the several projects in progress, the challenge is to obtain coherent data and to fully understand it. The second is to provide a customer experience that is more fluid and personalized. Fashion brands are at the very beginning of the journey. The third is to nourish creativity, generate ideas and facilitate the decision-making process (pricing, assortment...).
For marketing, big data provides an extremely fast test & learn approach according to sales, adds Benoît Sabatier. In the luxury segment, specifies Edouard Fonkenell, data allows us to explore and to create experience-based shopping sessions. It can also help optimize an economic model that is largely based on it products: how can we identify them early on and promote them quickly, so as to avoid losing energy on other products?
In the mass market, big data must help brands optimize the long tail: those products that are not among our best sales, explains François Nguyen. Should they be discontinued or kept in the catalog? Will some be tomorrow’s best sales? "Thanks to technologies like deep learning, it is possible to detect weak signs that are indications of early trends," he adds.
What are the pitfalls of big data?
"An algorithm is never perfect, there are always errors!", acknowledges Edouard Fonkenell. "But the objective of using data is to perform a bit better than if we did without. Progressively, we will succeed – even if the information is fairly fragmented – in gaining insights, in extracting interesting value".
The potential of big data is huge, as long as we know how best to use it, grants Nicolas Santi-Weil. We need complex algorithms to tap into wealth of data. The solutions that work are those that are very effective for the company but very simple for the consumer. The startup Easysize, which helps online buyers find the right size, is a good example.
For François Nguyen, now that we fully understand the consumer path on an e-commerce site, we need to tie this customer insight in with intangible product data (images, styles).
Is big data set to turn the entire fashion industry upside down?
First, data has the ability to place the consumer at the center of the creation process, as an influencer and trend setter. "The transformation of the fashion industry, with the appearance of data scientists, is fascinating," says Benoît Sabatier. "Big data makes the entire industry more exciting, without taking away the magic that is inherent in fashion".
Thanks to upcoming algorithms, fashion will need to reinvent itself as it will be increasingly determined by data, explains François Nguyen. This will be a challenge, not only from a technology perspective (such as finding a killer app to facilitate micro-decisions), but also from a human, transformation standpoint.
The impact will reverberate up to physical points of sale, predicts Edouard Fonkenell. In the luxury segment, shops will be the place where qualitative data can be collected. Until now, vendors have been writing down customer information in their personal notebooks. From now on, capturing data will become far more professional.
Is the fashion market late in utilizing big data?
For Nicolas Santi-Weil, the current debate over big data versus creativity is similar to the tension that existed between luxury and marketing ten years ago. Everybody now understands that they are not conflicting, and that big data is bound to become an industry staple. If the fashion industry appears reluctant to take on big data and has been lagging behind, the gap will be bridged once it has had time to adjust to these new technologies."
"Compared to markets such as automobile, all fashion companies are late," observes Maximilien Abadie. This comes from the fear of change and the concern that creativity might become dominated by data. It is true that the absence of standards in the fashion industry makes it impossible to set up product data history and makes data analysis more complicated. The major players in the fashion industry should however be able to standardize their processes enough to launch large big-data-based transformation projects across their organizations.
"The fashion industry underestimates the volume of data it sits on, and what it can do with it," says François Nguyen. Once they have understood the complexity of the task ahead of them, fashion companies will need to significantly invest in technology, people and processes – as already have telecoms operators, banks and insurance companies.
And if the fashion industry decides to simply ignore big data...?
If fashion companies do not embrace big data, others will, at their expense. Amazon has major ambitions and could well become the Netflix of fashion, warns Nicolas Santi-Weil. They are set to create it products using the same methods as in the House of Cards TV series... and we all know with what success!
"Data is necessary but not sufficient," he tones down. The power of GAFA* is scary, as they know how to respond intelligently to demand, but data will never replace creative genius. Fashion cannot be normalized. "Paris has a role to play: to create something that no one needs today but that everyone will want tomorrow!".
"The GAFA can destabilize the market by capturing the value upstream, but creativity is difficult to replicate and this is what protects fashion," adds Edouard Fonkenell. For François Nguyen, fashion brands can fight the battle through customization. "Amazon can succeed in selling basic garments, and they will do that very well. But there needs to be an artist behind certain types of clothes. Besides, building a heritage in fashion takes time and requires risk taking," concludes Maximilien Abadie.
* GAFA: Google, Apple, Facebook, Amazon