Is it possible to introduce big data solutions to an industry that is based mainly on intuition, unconventional thinking and sixth sense?
Does the world where clothes become unfashionable at the moment when a new blockbuster is produced or when a pop star’s big tour begins allow for the use of algorithms and machine learning to design something extraordinary?
Admittedly affirmative answer proves the fact that fierce competition and rapidly changing market trends force the application of new technologies that can assist fashion designers in their work.
This thesis is confirmed by a study conducted by the scientists from Penn State University. Analyzing the keywords, they managed to identify the source of ideas and works of influential designers, the affected areas and trends considered to be exceptionally far-reaching. Considerable amount of data from various sources, especially from Internet portals such as Instagram, Pinterest, Twitter or Facebook, makes it possible to determine the correlations and emerging tendencies and to search for the models that can be applied to the fashion industry. The aforementioned study was based on opinions from a renowned Style.com portal, former website of Vogue – one of the most esteemed fashion magazines. 6629 opinions provided by 816 designers in the period between 2000 and 2014 served as a basis for the analysis of 30 winter-summer seasons.
Professor Heng Xu’s team extracted the keywords related to the style, color, fabric and many more different details observed in the collections of each designer. Next, the scientists developed a model referring to the experience, rank and area of influence of particular designers. In order to enhance the accuracy of the model, the scientists contrasted it with three renowned fashion rankings: “Times”, “Fashion Merchandising Degrees” and “Celebration of the 20 Most Influential Designers”. It turned out to be surprisingly precise.
In most cases, the attitude of specialists from various fields towards data analysis is generally positive. It should not come as a surprise though, that the industry that perceives itself as a sophisticated form of art quickly criticized the model described above. Skeptical fashion designers did not accept the results of the analysis, claiming that in this particular case such models cannot be properly evaluated.
Analyses performed on the delivered data allowed the scientists to conclusively prove that big data solutions become a must-have for the industry. They do not only support the designers’ work, allowing for the performance of predictive analyses or identification of the area and source of the trends, but also make it possible for the beginning designers, whose future works will be sold by the most renowned fashion houses, to develop their careers in quicker and more efficient way.
On the other hand, the technology will help the clients to effectively plan their wardrobes and to adjust them to their style and funds. Works produced by successful designers are expensive and unavailable for most people, but if information about an oncoming trend is provided quickly enough, fashionable clothes may turn out to be affordable for them as well. Moreover, just like it is in the case of history, fashion trends repeat themselves, enabling the prediction and use of currently fashionable clothes in the future. Perhaps, even to a small extent, the minimalism based on predictive analysis will influence the contemporary consumptive tendencies by reducing the amount of clothes being thrown out or sold to thrift stores.
Interestingly enough, results of the studies conducted by the scientists from New Zealand are confirmed by the invisible hand of the free market and can be illustrated with the example of companies that analyze large amounts of fashion industry data, such as Editd or WGSN, which increase their revenues very rapidly.
During the recent years, the London branch of Editd has collected 53 billion points of data ranging from garments for men and children to women’s accessories. The set includes data from over 1000 retail stores and 15 million photos from all over the world. We need to highlight the fact that the data came from many various sources, such as Internet sites, social media, industry reports or blogosphere, and it is processed in real time. Such an approach caused a decrease in the amount of discounted products, for each such product is actually a failure of its designer and distributing store. The company uses this information to perform analyses and develop reports that can be used by the designers to adjust their products to the changeable market conditions. Moreover, if we are to believe the sources, due to the analytical data provided by Editd, Asos – one of the leading eCommerce stores in the industry, improved its sales by 33%. Paradoxically, the whole thing may also have a considerable impact on… medicine. Size of the delivered clothes may be used to analyze and predict the diseases related to obesity and anorexia. It may sound incredibly, but the first models that can be applied in this manner are already being developed.
WGSN is another successful fashion company that created a tool which enables the prediction of future trends. Here, the amount of collected data is slightly lower than in the case of Editd: the set covers over a million products and 11 million SKU from over 10 thousand eCommerce stores in the whole world. Still impressive. Apart from the performance of predictive analyses, the tool makes it possible to lower the costs through the supply chain optimization. Data analysis gives answers to the key questions: which collections should be produced more intensely, how many resources are needed, do we need to store the goods locally or have them delivered on demand, etc. Technological development will cause an increase in the amount of applications, even though we have lots of them even today.
To conclude, it is most likely that big data solutions will not redefine the fashion industry. Too many elements will still depend on human features such as creativity, innovation or genius. However, the world of numbers, data and algorithms will significantly influence the development of this area, shaping its frames and creating new, unknown possibilities.
Quoting the COE of Editd – Geoff Watts: “We help retailers have the right product at the right price and the right time. That’s the kingmaking thing in retail. When you get that right, it unlocks a fortune.”
Are you also passionate about big data and want to learn how to use it to improve your e-business? Write to Divante.