Fashion is a high-risk business. For luxury retailers, as well as high street players who take their lead from the runway shows, financial success means getting a handle on which designers, collections, trends and ‘it’ pieces will sell best. Take a moment’s rest and you stand to lose out.
For this reason, the buying decision is crucial. How much of what, where, should retailers stock? Market research, as well as following what fashion editors and influencers have declared popular, is often used to guide this process.
However, these indicators are fallible, as consumer behaviour can be hard to predict. Trends are dictated by myriad higher powers, and if one spoke of the wheel snaps – say, for example, an eco-conscious brand is outed as being environmentally irresponsible; or a designer is found to be exploiting minimum wage workers or being culturally insensitive to a key market – then it can all come tumbling down.
What every retailer wants is as much certainty as possible; increasingly, that means harnessing the power of cold, hard data.
And where does most of this data come from? Why, the place where conversation is often more revealing than in person: social media.
Machine learning is an integral part of modern-day business – countless processes are iterated and improved as artificial intelligence is applied to discover insights and deep-seated truths. Applying the power of AI to social media is a no-brainer, really – scraping the likes of Twitter, Facebook Instagram et al. can deliver critical business insights when it comes to the key Fashion Weeks and what creates a buzz.
Because at the moment, the long and short of it is designers don’t know how many pieces they’ll produce in advance – they usually refer to previous sales data from similar items, in conjunction with specific stockists and regions. They (quite sensibly) secure safe bets.
Going to social media platforms to crunch the data and see exactly which collections, pieces and trends generate the most positive buzz can allow fashion retailers to plan buying and marketing strategies with much more precision. When it comes to ordering X amount of units for specific markets, it could easily make the difference between profit and loss.
Circling back to the designers themselves, similar social and web data derived from online chatter around the shows could also be used to better inform how much they need to manufacture, how likely a certain item is to sell and so on.
Some brands, for example Gucci, are already harnessing the power of such insights to inform their collections. Lanvin also gauges reactions from the public and prominent industry figures this way, pinpointing which looks are most popular. It used this method to measure the effect of its recent relaunch too, digging into how people recognised and interpreted its new brand ethos.
The traditional seasonal collections still exist because, for the most part, retail buyers need to view collections and decide which they want to go with; and designers then need to process and manufacture these orders. But that calendar is becoming increasingly disrupted and timescales compacted to deliver stock to much shorter lead times. This trend will only grow as machine learning is used more frequently to inform key manufacturing and buying decisions.
Of course, print media still requires a lengthy lead time to photograph samples for publication, but that aside, there’s not much in the process that couldn’t be accelerated applying machine learning to scrape social for key insights.
It’s not a quick fix, nor is it the answer to everything. But fashion is dependent on trends and customer sentiment. In a fast-moving digital world, every industry needs to do what it can to stay ahead. That is why machine must meet couture – it’s the only way luxury fashion can truly keep pace with its audience.
First published by Luxury Daily