Home Featured Analytics, AI & ML: Accelerating PLM Across Fashion’s Value Chain

Analytics, AI & ML: Accelerating PLM Across Fashion’s Value Chain

0

In his first piece for 2021, Founder & CEO of WhichPLM Mark Harrop, shares the benefits of using Artificial Intelligence & Machine Learning within PLM. He believes that if PLM vendors accelerate and expand their use of AI it will become, without a doubt, a critical element of any fashion company’s 2021 strategy.

During the last 25 years, the fashion industry has turned from one of pure bricks & mortar to one dominated by a group of goliath e-commerce businesses – think Amazon, Jingdong (JD.com), Alibaba, eBay, Rakuten, Zalando, and so on. Beyond these we also have a vast growing number of businesses that either started as pure online companies or transformed from B&M (bricks & mortar) to become pure play online businesses. As retailers scramble to survive the online onslaught, we continue to see an increasing mixture of both the traditional B&M and the online worlds coming together. It’s these sector groups that will unquestionably weather the storms.

“The battle to survive and prosper is well underway and has been accelerated by the dramatic effects of the pandemic, and now the underdogs are scrambling to join the elite on-liners”

The challengers will not find it easy to gain new customers in the online world and it’s crucial that they think out of the box and totally reimagine fashion’s ever changing end-to-end value-chain age!

Based upon my 4+ decades of experience working at every end of the supply-chain, my strong opinion is that fashion retail (B&M & online) supply chains are totally disconnected, apart from, in the main, a contract or purchase order shared with manufacturers (electronic or paper form) that are used to place orders for products.

It’s only in recent years that we have started to see PLM solutions moving to a near real-time collaborative platform for design & development purposes, and even in these cases there is still only approximately 2,000+ PLM implementations across the entire fashion industry (across all fashion sectors). In 2021, the industry needs to make some pivotal, bold steps to take PLM beyond design, sampling & development and expand its usefulness into the extended value-chain to support the contract and P.O. workflow processes.

“It’s time for a dramatic shift if the goliaths of the fashion world are going to be challenged or even if fashion businesses just want to keep pace”

This will require a new way of thinking that will need to stretch from trend research, design & development, thorough manufacturing, logistics, warehousing, retail, or direct to consumer and even beyond to disposal of products.

What are some of the critical tasks that will help to streamline the end-to-end workflow?

Obviously, critical is the speed and accuracy of interpreting consumer behaviours. Retailers and brands will need to mine vast amounts of rich data in near real-time, analyse patterns and insights which will feed machine learning (ML) algorithms to help support merchandising, buying and design decision making. Artificial Intelligence (AI) and ML are already playing a crucial role in helping retailers deliver stellar customer experiences across different fashion channels. For those new to AI, it’s worth mentioning up front that AI comes in many forms such as: computational intelligence, informatics, knowledge-based systems, and cognitive systems.

I’ll use the catch all umbrella term of AI for the rest of this article, so, let’s look at some of the opportunities that are already being delivered via AI and what we can expect in the not-so-distant future, especially as it relates to PLM.

Knowing your customer (KYC)

Today, AI enables retailers and brands to understand day-to-day or even minute-by-minute changes in consumer behaviour linked to every interaction with a given business. It empowers the first movers of AI to filter through hundreds of parameters such as our likes & dislikes, our favourites, past buying patterns, gender type, location, interests, sizing data, returns data, and spending power and then uses the resulting data-insights to provide suitable product recommendations to each customer.

Ultimately the goal is to predict a consumer’s behaviour and create a positive experience that leads to new sales. To be able to accomplish this task, a prerequisite will be the use of large amounts of data. In a nutshell, today’s smartest retail AI implementations are plainly, prediction machines that use algorithms (a process or set of rules to be followed in calculations or other problem-solving operations, by a computer) to analyse large datasets, in order to optimise the goal of selling fashion. As they perform the optimisations, buyers and merchandisers continually learn and tweak the algorithms to improve their accuracy (hit rates) and recommendations.

The best thing about AI algorithms is that they are designed by humans that are continuously learning from the evolving patterns and insights.

How might we benefit from using AI together with PLM and best-of-breed design & development and manufacturing solutions? There are many potential options for using AI in fashion.

We can use the trend and demand inputs coming from AI to balance the throughput of our value-chain partners. We can use the same AI algorithms to not only accurately predict trends, but also to use visual search engines to turn these trends into design assets. Beyond the design stage we can continuously use the real-time AI feed to affect future demand and enable manufacturing, e-commerce and physical stores to plan accordingly.

AI also enables visual search, an emerging development that is already being used in retail’s downstream marketing processes, as well as already being used as part of the expanding PLM platforms that are now supporting product design & development. It has the potential to revolutionise how designers and developers will be able to search for image assets (products, colourways, patterned materials, accessories etc.)

Another exciting use-case is AI search and compare capabilities, which will enable a user to request a search of the PLM database to find and compare products or materials that are similar to their visual example. This process will enable users to move beyond text-only searches to the use of visual search, providing substantial efficiency benefits for all those concerned.

AI will also find its way into product development; once designers have developed the basic design requirements, technical developers and garment technologists will be able to ask AI to search for similar products that matched the tagged images, main product types, specialty, target country, sizing details, etc. From this point, we can expect PLM to automatically build a basic specification (template style) that offers the nearest match to the designer’s request, including the base data and process elements, circumventing the need for each person to enter or replicate the entire dataset.

We will be able to use AI-based algorithms to provide valuable knowledge that guarantees stability in development across our value-chains, improving transparency, operating together with IoT (connected devices), factory planning & monitoring, all the while lowering costs and reducing margin risk for everyone involved. It goes without saying that manufacturers will need to up their game and deploy new technologies (PLM, Factory Planning, Synthetic Costing, IoT & AI platforms), these transformational upgrades will help to support predictive analytics, smooth out maintenance and downtimes, improve efficiency & quality, and ultimately these new AI-powered tools will help to optimise their end-to-end value chains.

As both the downstream retailers and upstream manufacturers start to deploy AI into fashion’s workflow processes, there will be a need to educate employees at all levels across the value-chain on how to read, question, understand, and interconnect with the resulting data.

Using AI within PLM, we will spend less time building and sending Tech-Packs, and more time focusing on making accurate data-informed decisions.

AI should be seen as our new personal assistant

Fashion retailers will have to develop the most optimum mix of AI and human work. Retailers, brands & manufacturers that want to get ahead of the game, will need to work as a single entity (an infinite data loop, if you like) using PLM & AI with other best-of-breed solutions to help deliver the most appropriate fashion products based on relevant and real-time customer data. The fashion industry will become far more adaptable with the ability to react according to ever-changing demands and will also be better positioned to predict future requirements.

It is, in my opinion, critical that PLM vendors should work together with their fashion retail customers and value-chain partners, to accelerate their thinking toward the broader deployment of AI across the value-chain. If PLM vendors accelerate and expand the use of AI, it will become without any doubt a critical element of any fashion retailer’s 2021 strategy. The digital transformation of businesses must not only digitalise the downstream product design and development processes, but it must now transform the roadmap to include the entire value-chain, using AI-level analytics to support an understanding of the where, what, why, and when of each task. It will help us understand how we might use the resulting data insights to improve how something happened, and the potential for further improvements as part of advanced decision-making processes.

Mark Harrop Mark Harrop is the founder and Managing Director of WhichPLM. During a career that has spanned more than four decades, Mark has worked tirelessly to further the cause of PLM – providing the unbiased, expert advice that has enabled some of the world’s best known retailers, brands and manufacturers achieve efficiency savings across their entire supply chain through informed technology investments.