Fighting fire with data
By Adam Driussi
British fashion giant Burberry found its brand under fire in recent months – both literally and figuratively – after burning more than $50 million worth of unsold luxury goods. This is not an uncommon practice, with many high-end brands preferring to destroy stock rather than risk it falling into the hands of discounters.
However, in an era of keen corporate social responsibility, these kinds of practices can only carry on for so long before a company is called out for wasteful behaviour. In this case, Burberry has issued a heartfelt mea culpa and promised to expand its efforts to reuse, repair, donate or recycle its unsaleable products.
Recycling is becoming a cause célèbre in the fashion world, led by charitable organisations such as the Make Fashion Circular program, which is run by the Ellen Macarthur Foundation and boasts the likes of Burberry, Gap and Nike amongst its partners. Unfortunately, while the raw materials from unsold goods can be recycled, there is no way to recover the time, effort and resources such as water and energy that go into their manufacture.
A more effective solution is to better forecast demand to avoid surplus stock in the first place. At Quantium, we’ve been using consumer and retail data to provide solutions like this for well over a decade, but traditionally this kind of analysis has been more at home in fast-moving consumer goods, where velocity of sales is high, and consumption tends to be repetitive. How would it translate to the notoriously unpredictable world of high fashion?
Wherever there is human activity we can find traces of behaviour, in the form of data, to enable this type of decision-making. With the exception of true haute couture, most luxury items are produced in significant quantities and can be placed into discrete product categories – say trench coats, or men’s fragrances. Sales data exist for each of these categories and the impact on sales of factors such as pricing, brand health, marketing initiatives, geography and the presence of competitors can all be measured.
But it’s not that simple – the very word “fashion” implies the enormous significance of subjective opinion on the performance of an item. That too can be measured, whether by analysis of positive vs negative product reviews or uptake by influencers. The advent of artificial intelligence allows us to conduct this kind of analysis with a scale and sophistication undreamt of a decade ago.
A decision engine using modern machine learning techniques driven by these inputs could create an early-warning system, which would allow the manufacturer to quickly determine the likely continuing pattern of sales and adjust their production accordingly.
Is this feasible for a company like Burberry? It should be, but it all depends on its data – its availability and readiness for analysis. Starting from scratch, a project of this scale could take two to three years to properly implement. With the right data sets it could be a matter of months.
Even in industries like fashion, which are highly temperamental and subject to many external factors, data and machine learning can be used to execute commercial strategy – such as protecting a luxury brand – in a way that is better for a firm’s reputation, better for its bottom line and better for the planet.
And while it’s understandable that high fashion might be slow to join the data revolution, the same excuse can’t be made for retailers in other sectors where the information is already at their fingertips. Take mainstream fashion, where sales volumes for a line of jeans can quickly generate a trove of data, and where trends often move much more slowly than on the catwalks of Paris and New York, allowing much more accurate use of historical data in predicting future performance. Waste still occurs in this space, even if it’s more often found in steep discounts than in the bottom of an incinerator.
The technology and the data expertise already exist to make this kind of inefficiency a thing of the past. To thrive in a competitive global environment both brand owners and retailers must take full advantage of it.
Whether you are just starting out or advanced on your data journey, we’d be delighted to speak with you about the ways in which data science and AI could be transformational for your organisation.