How data can address the fashion industry’s tech problems

11 Nov 2021

Image: © Gorodenkoff/Stock.adobe.com

While many industries are digitising, the fashion industry has been a bit slower to adapt. However, change is on the horizon with data opening up new doors.

Technology has transformed the world of shopping and fashion. E-commerce and payment tech has made shopping easier, with millions of products available within just a few clicks, and machine learning algorithms are improving how we browse for things online.

But the fashion-tech industry faces challenges, particularly when it comes to subjective nature of fashion. While the technology and the data might be there, how to best use these in the fashion space can be up for discussion.

‘It’s easy to see a world where shopping journeys begin rather than end with the shopper’
– EOIN CAMBAY

Eoin Cambay is the CEO and founder of Swan, a Dublin company that has built a sizing recommendation tool for fashion retailers using artificial intelligence. Cambay said that while size is objective, it’s when companies start trying to bring AI to subjective areas such as style that the problems start.

“[Style and taste] are largely driven by people you admire and spend time with, be it online or in person. This is likely a reason why online styling services have struggled to win most people over – they’re using the wrong data. They’re based on price points, styling quizzes and what’s available,” he said.

“We need a shopping experience that is personal and truly caters to the needs of shoppers. Personalisation that is genuine, not a first name at the beginning of an email or a financial Band-Aid like free returns.”

Cambay said rather than overcomplicate fashion technology by attempting to create algorithms to guess at a shopper’s subjective style, technology should be created to answer the questions that shoppers already have.

“It’s easy to see a world where shopping journeys begin rather than end with the shopper. Clothes are made to order from shopper measurements, preferences are accounted for with customisation, and reviewed in 3D. No more waste or standardisation.”

Utilising data

While subjectivity is an understandable challenge facing the fashion industry, there is still a need to embrace digitisation. After all, data could unlock new opportunities for retailers, brands and customers, if leveraged correctly.

Eamonn Donlyn is the CEO of Darwin & Goliath, a Trinity College Dublin spin-out based at the Adapt research centre, which provides a product recommendation engine using 120m data points.

“Algorithms are only as good as the data you feed them and sometimes that limits their effectiveness,” he said. “Many solutions in fashion rely on a single algorithm to solve a problem, such as product recommendations at the bottom of a product page on a fashion retail website.

“This one algorithm model is very inefficient and leads retailers to place a ton of widgets around a site – to add more algorithms – bombarding customers with overwhelming experiences.”

‘People are happy to share information if you engage with them on their own terms’
 – EAMONN DONLYN

Donlyn echoed Cambay’s sentiments that in order for the fashion industry to digitise successfully, it must address shoppers’ actual needs rather than what it thinks they need.

“Consumers are demanding products that are personalised, customised, ethically and sustainably produced and delivered rapidly,” he said.

“In order to properly serve these trends, you need customer data so you can react to these consumer demands properly. The approach of ‘design-make-sell’ is dying and incumbents are at risk to those that move toward a demand-driven model with direct-to-consumer rising.”

He said that product recommendation tools and outfit sizing platforms such as Swan are a great start for the sector, but that fashion companies need more natural ways to interact with customers to understand their needs.

“While loyalty systems are meant to be a key method, 77pc of them fail within two years. This is because they only reward purchases, which does not provide enough value for customers to engage when they aren’t buying something,” he said. “People are happy to share information if you engage with them on their own terms.”

Donlyn said that while there are plenty of tools available that can help fashion companies use their data, these tools are often seen as something a company cannot afford. “But if they wait too long to change, they will suffer a similar fate of mall owners and taxi companies have faced in recent years.”

Updated, 10.49am, 12 November 2021: This article has been amended as the company Syze has rebranded as Swan.

Jenny Darmody is the deputy editor of Silicon Republic

editorial@siliconrepublic.com