Big data is proving to be a boon for fashion forecasting.
Fashion forecasters claim to tell fashion’s future. But are they prophets: intuitively reading the world of today to see the world of tomorrow, or are they dictators of taste: telling companies what colours to pick, and watching as their ‘predictions’ become a self-fulfilling prophesy?
I’ve always thought fashion forecasting to be the latter. Sure, they know what’s trending, what’s worrying, what’s inspiring. Their research tells them whether consumers want to spend big or indulge in fleeting sensory experiences or cocoon themselves away from the world.
But all this knowledge is hard to quantify, and I’m not convinced that they can accurately coalesce it into the specifics of what colours people will want to wear in eighteen months. Rather, they make an educated guess, then sell that information at a premium to textile mills, designers, companies… who then produce cloth and shoes and jackets in that colour, making the forecast come true.
Close to season forecasting seems more straightforward: look at what’s happening on the catwalks and make connections to propose ‘trends’. Style.com has this down to an art form.
But now big data is disrupting the forecasting process. The use of data analytics in the fashion industry is about mining fashion’s present to find fashion’s near-future: a liminal space in time between the now and the nearly-now. Let’s call it ‘the fashion presently’.
People create sets on Polyvore, ‘like’ items on online stores, pin garments on Pinterest and add items to their Stylitics or Clothia wardrobe. All the while, data-bots are channelling, filtering and analysing this activity.
All this data is synthesised into reports itemising and graphing how fashion garments, colours, styles, brands and other variables rise and fall in consumer favour. For fashion designers, if you can see a colour, or style, or textile print making it’s way up the charts, then you can rush it to market before it hits its apogee a touch after ‘the fashion presently’.
Editd is a business to business real-time analytics service. But the others are consumer-facing social networks or apps. We are often told that if an online service is free, it’s because we’re the product. This is certainly true for Polyvore and Stylitics. Stylitics is especially brilliant. You give it everything in your closet: itemised by brand, photographed, with data on its size, colour, where you bought it, how it fits and how often you wear it. Or in the case of Polyvore, you assemble the items you covet into a shoppable collage. And you do these tasks voluntarily, for fun.
But this information is incalculably valuable to brands, provided of course that there are enough users to make the data meaningful. And there are. Polyvore has twenty million active users. Stylitics is relatively new, but is growing strongly. And unlike Polyvore, Stylitics was built from the ground up as a data analytics tool for the fashion industry. It’s just masquerading as a social network for sharing one’s closet with the world.
So fashion forecasting is having a kind of Enlightenment moment – moving from the hazy realm of prophesy to the science of data mining. But even so this data won’t tell you the fashion future, just the fashion presently.