What’s fashionable in the world of data?

At the end of Paris fashion week we look at the unexpected yet fascinating role of data analytics in the fashion industry.

As one of the most important events in the fashion calendar, Paris Fashion Week, comes to a close, the catwalks remind us just how much we revere the creative, the inspirational, the artistic and the fluid. But peek beneath the surface and you’ll find a story about something less glamorous perhaps - data – which increasingly plays an unexpected yet fascinating role in the fashion industry.

Like cars, memory speeds and the decline of sneaker wedges, fashion has always been getting faster. The invention of engraved fashion plates in magazines accelerated fashion trends in the early 1800s. In the same way, today’s digital technology harnessed to global cultural capital and influence has led to a fashion cycle that now revolves at breakneck speed.

So let’s stop for a moment to take a look at three important data themes trending in the world of fashion right now: profit, personalisation and prediction.

Profit

We’ll tackle profit first. Fashion retail – both in-store and online - is a complex operation and one with good potential to exploit data to improve the bottom line. The allocation of stock to stores or warehouses, transportation options, management of inventory, supply, distribution, sales forecasting, optimised price points, and database cleaning can, to put it simply, all be made much better through data analysis.

Not without challenges, of course. For one thing, there’s just so much information in different formats: there’s all the stuff collected from customers through social media, shopping carts, surveys and loyalty cards, not to mention at point of sale through SKUs (unique product tags) and payment cards. None of this is helped by the probable mismatch between supplier and retailer IT systems along a convoluted international supply chain. It’s a business which has worked on historically slim margins - and without many boffins - so there may not have been a great deal of investment in the IT suite.

All that said, thirty years ago stores had no real way of knowing which sizes were in demand: if a browsing shopper couldn’t find their preferred size they would simply not buy, and knowledge of that unfulfilled order was lost for ever. So when it comes to data use there are definitely some easy wins for both buyer and seller.

Personalisation

Even quite recent predictions expected us to be living in a hyper-personalised world where not only would clothes be sized and made exactly for us, but our online engagement, search terms, posts and browsing habits would be analysed in real time to curate a completely personal shopping experience. This would follow us into store where eye-tracking and behavioural analysis would pre-empt and direct our every move, making us unlikely to leave empty-handed.

Data protection and privacy laws, particularly on biometric and location data, and the reach of personalised advertising, mean that this type of super-curation may in fact be a limited edition. Personalised pricing is heavily dampened in Europe by rules about competition, consumer protection and discrimination.

One way to bypass some of these issues is through loyalty schemes, which have popped up everywhere. To the converts, these are ways for brand-loyal customers to be rewarded and further engage with favourite products. But to the sceptics, it’s effectively making privacy a luxury add-on: and for this reason we have also seen questions asked about supermarket loyalty cards, and about generalised data collection at scale, in more than one EU country. So the prediction is that extreme personalisation may be too challenging. Which brings us to...

Prediction

As we’ve said, cultural information about what’s going on in the world has always massively influenced fashion. When information is harnessed to very new data technology we see, for example, the use of AI to predict fashion trends by analysing social media images. Or along similar lines, the generation of “sentiment data” (“customer feedback” in old-speak) based on social media keywords, or voice tone analysis.

But, in an infinitely connected world, when we use the same algorithms to identify trending products and make them available really fast, does this mean we all continue to end up wearing the same thing? Could fashion ever be reduced to a formula?

Frustratingly for the forecasters, fashion seems to follow the same rules as chaos theory: that complex systems react strongly to unexpected stimulus. And there’s nothing that fashion likes more than a disrupter: a kind of Mary Quant-um physics, perhaps.

And to come?

On a more serious note, fashion retail has major challenges ahead. Among them are global economic headwinds, changing attitudes to consumption, and endemic oversupply in the wake of our continued love affair with fast fashion (millions of tonnes of clothes go straight to landfill every year). 

In response, we are beginning to see some extremely creative new use of data: such as tools to get sizing and styling absolutely right and predict likely purchases with pinpoint accuracy, to minimise returns. Not to mention the possibility of shifting the whole fashion circus – including the clothes we “wear” - into the metaverse instead.

It seems that for the moment at least, fashion and data analytics are a pretty perfect fit.