Today, more than ever, social media – and Instagram, more importantly – is undoubtedly the luxury industry’s pot of gold waiting at the end of the rainbow. And Heuritech, a Paris-based company, intends to turn this pot of gold into valuable nuggets of market data.
In fact, the company just released the Spring/Summer 2020 main industry trends across all four cities (Paris, Milan, London, and New York) by blending data-driven social media insights and human analysis into an artificial intelligence tool that predicts the behavior of trends and the performance of key products.
When it comes to the SS20 fashion week performance, the reach of New York Fashion Week was the highest – 39% higher than Paris Fashion Week’s reach. “Despite the lower number of shows in New York, fashion week in the Big Apple remains highly attractive and gathers very influential figures: the average reach of NYFW is the highest of all shows,” Célia Poncelin, Heuritech’s Head of Marketing & Communications, stated, noting that the sharing of Instagram posts this season was still led by PFW (36%), followed by NYFW (31%), MFW (22%), and finally LFW (11%).
As far as the “top brands” category is concerned, the five best-performing brands for the SS20 season based on the sharing of posts and brand mentions on Instagram include Chanel (17%), Dior (13%), Gucci (11%), Prada (9%), and Versace (7%). The latter entered the top 5 this season by breaking the Internet with Jennifer Lopez, who wore the iconic jungle dress for Versace’s runway finale. “Legend has it that Jennifer’s dress made such an outstanding impression during the 2000 Grammy Awards, and caused so many image searches on the web, that the actual invention of Google Images was determined by this milestone moment,” Poncelin added.
The “top designers” category of the SS20 season were led by Karl Lagerfeld, who was followed by Anthony Vaccarello, Maria Grazia Chiuri, and Demna Gvasalia. “Karl Lagerfeld was by far the most mentioned designer, a testament to how much the industry is still paying him tribute. Vaccarello was also in the spotlight for sublimating Saint Laurent’s iconic women’s suit,” Poncelin continued. “Mentions of Maria Grazia Chiuri mostly highlighted the Dior garden as a way to address the sustainability issue. Demna Gvasalia mentions focused on his departure from Vetements as well his SS20 show for Balenciaga with its bold European Union blue set and diverse model cast.” As for the “top emerging brands,” Heuritech’s report showed a clear hype around four independent designer labels: “Spring/Summer 2020 was marked by independent brands whose names stood out in fashion week-related posts: Marine Serre, Sies Marjan, Telfar, and Richard Quinn made it to the top of the ranking of new fashion brands, just after Jacquemus – an absentee of the SS20 season of PFW, who was nevertheless very present on social media.”
Heuritech’s numbers are less promising, however, when referring to the luxury industry’s current battle horse, namely sustainability. “While consumers are embracing the sustainable trend, fashion week posts don’t reflect this engagement,” stated Poncelin, noting that only 1% of posts from fashion week mention sustainability. That being said, the #sustainablefashion hashtag has grown fivefold in the last three years, and sustainable brands attract more user awareness than ever before – 50% of posts about Stella McCartney’s SS20 show mention “sustainability” in the caption. When it comes to macro-trends, such as colors and textures, Heuritech’s visual recognition technology allowed the company to quantify the palette of the SS20 season by calculating shares of every color and texture, which ultimately indicates the power they will have. The best-performing colors for SS20 include off-white, beige, bright pink, rust, and purple. As for the textures, the company has set its focus on bags, noting that animal patterns, such as snake and cow, as well as pearl embellishments, padded and quilted fabrics, as well as fleece, were the best-performing textures of the SS20 season.
As of today, Heuritech intends to firmly establish itself while competing with already existing trend forecasting agencies (WGSN and Nelly Rodi, amongst others) and AI-technologies (IBM Watson and Antavo, amongst others) dedicated to the luxury, fashion, and lifestyle sectors. The company is currently working with luxury brands, such as Louis Vuitton, Dior, and Adidas, and supporting brands in their decision-making, notably in optimizing product assortments, forecasting product quantities, and mastering product positioning and marketing strategies.
“For this Spring/Summer 2020 fashion season specifically, we analyzed over 40,000 images related to all four fashion cities, which resulted in over 2000 detected trends,” Poncelin explained. “By capturing early signals from our ‘edgy’ audience panel on Instagram, we are able to forecast expected growth from the fashion forward, to the mass. The breadth and depth of our data science analysis allow brands to forecast more accurately throughout their product’s lifecycle while predicting trends and avoiding overstock.”
When referring to the so-called “edgy” audience panel, Célia Poncelin referred to the Instagram content shared by a panel consisting of the industry’s top-performing professionals and influencers mostly cherry-picked from the BoF 500 list – but not only. "The goal is really to be representative and neutral about the people who make today's fashion industry. We do manual curating with fashion experts but also automatic curating, by looking at 5 years of historical data," explained Poncelin. Their voices on Instagram have served as a basis for the company’s latest report. More precisely, the report’s results have been generated by an artificial intelligence-powered image recognition tool that predicts what product trends are coming and how they will behave using real-life pictures on Instagram and Weibo. This image recognition tool analyzes a myriad of aspects, such as colors, styles, textures, patterns, shapes, logos, but also specific SKUs when available. Over three million pictures are, in fact, analyzed each day, resulting in over 4000 detected trends and products with a 90% precision rate, according to official company statements.
However, there are some limits to this “picture-perfect” Instagram-based forecasting and its ultimate relevance when predicting future industry trends. As of today, the nature of the relationship between brands and top influencers – whose posts are included in a forecasting analysis – is rarely an organic match, and often a transactional one. Paid product placements are not always unveiled as “sponsored posts” on Instagram. In fact, whether posts are labeled as advertising (#ad) or not, is often left to the influencer’s personal choice. This constitutes a challenge, and ultimately means that social media trend forecasters need to ensure that their “top trends” are based on a performance that is organic, transparent, and not compromised by commercial agreements, if they want to deliver the most objective market data to their clients. A challenge that they still grapple with.