NATIONAL: Cue, Saba, Aurelio Costarella and Antipodium were the most talked about Australian fashion labels in social media last month, according to a new trend forecasting service.
London-based agency Stylesignal, founded by Australian expatriates Geoff Watts and Julia Fowler, has developed a web-based software as a service (SaaS) product which automatically analyses opinions on trends as they develop through media platforms such as Twitter.
Dubbed Trend Science, the system can break down key trends by region as well as garment types, prints, patterns, colours and detailing. Users are presented with a top 20 list for each category and can search for particular terms to see how opinion has changed over time.
For instance, a user who selects the search fields ‘womenswear’, ‘Australia’ and ‘label’ will see Cue ranked as the most talked about brand in October, followed by Saba, Aurelio Costarella and Antipodium. A monitoring function allows users to see developments within selected areas by setting alerts across blogs and Twitter sites.
Watts, who has a background in computer science and IT, said the technology had been developed over a two- year period.
“Speedy information helps with ordering or designing stock for next season,” he said. “As designers become more internet savvy, they are streaming their catwalk shows so now consumers can make a decision on the spot as to what they like and don’t like – and what to put on their ‘list to buy’ when the garments are in the shops six months later. Buyers and designers can utilise this information to make sure they have the right product in store.”
A patent is also pending on a feature which can quantify social media opinions on specific brands, styles and designers. Users simply type in a word or phrase and a specially generated chart illustrates whether opinion on the area is favourable or not.
Watts said the sources of opinion can also be traced, allowing brands the opportunity to address any genuine cricticism on garments or retail experiences.
“We monitor sentiment automatically using machine learning, natural language processing and a degree of manual curation,” he said. “We aggregate the data and allow our clients to inspect it using an easy-to-use interface that normalises the data into a standard layout.
“We automatically monitor about 12,000 curated fashion-centric sources that we update all the time and augment that with a general feed of ‘firehose’ tweets that we use to discover trends emerging from outside of our specific coverage.”
A search on the term ‘Rosemount Australian Fashion Week’ indicated Dion Lee, Aurelio Costarella and Arnsdorf were the most favourable labels on show. Meanwhile, vintage, punk, goth and futuristic were the most popular styles in Australia over the last quarter.
Watts said the recent spring/summer 2010 shows across London, Milan, New York and Paris recorded leather, denim and lace as the most popular fabrics and metal chains, sequins, architectural shapes and zippers as the most popular garment detailing.
The technology is offered as part of Stylesignal’s womenswear trend forecasting service.
Trend Science in action.
Assia Benmedjdoub