PLM & Fashion’s Future Demand Model

Prasham Kamdar, Managing Partner of Ptex Solutions, discusses the future of fashion’s traditional supply-chain model and how PLM solutions can (and will) adapt.

It’s no secret that the last few months have been bizarre, and concerning for many of us – especially in the fashion industry. And as we begin to come out of the global lockdown, businesses are going to need to rethink and reimagine fashion’s traditional supply-chain model.

It’s already clear to the vast majority of brands, retailers and their value-chain partners that our new reality will see lower consumer spending levels – at least for the next 12-24 months. Listening to the latest news and trends coming out of retailers and brands, its fairly obvious that consumer spending will be targeted more at ‘made to last’ sustainable products that will be linked to higher quality standards when it comes to design, fit, materials, trims & components, colours and manufacturing processes. Linked to these changes in consumer behaviour we can also expect to see smaller, more complex orders being driven by near-real-time consumer demand. And these types of changes are just the tip of the iceberg of what’s going to disrupt the fashion sector as we know it.

What we do know, is that PLM vendors will need to adapt their solutions to support the changes we’re already seeing; they will need to create third party APIs (application programming interfaces) to support value-added solutions that can help support these changes and investigate the use of new tools as part of a broader PLM eco-system.

So what can we expect?

Advanced Analytics using AI & ML

One area in which we’re already seeing benefits is in advanced analytics. And moving forward, Artificial Intelligence and Machine Learning can be used to better understand current demand curves, collecting near real-time insights and demands to support the design and development process, but at the same time understand the capabilities and risks of our connected value-chain partners.

Digital design & development, and getting to know your customer

When it comes to digital product design and development in the near-future, trend data and insights can be shared with your online storyboards, that can then expand this data flow to utilise links to material types, trims, components and their supporting meta-data residing on new material digital platforms. Several of these new platforms will include both virtual and real materials. Examples include: Adobe Substance for virtual fashion materials or real materials coming direct from a mill’s catalogue, delivered by digital libraries via new platforms; sourcing platforms and others that will be able to share data with PLM. What’s really becoming clear is that we are already moving ahead with more and more APIs between third party players.

Before utilising these new platforms to create products, you need to know what products you’re creating. Going forward it’s important, if not critical, that you really do get to know your customer – their likes, dislikes, actual shapes and sizes. It’s time to move away from the “one-size-fits-all” methodology. Today, we have the opportunity and new digital tools to measure our customers and it really doesn’t matter if you are a fast-fashion or everyday bricks and mortar business, we still have the same issues when it comes to return rates. Granted, not all returns are related to sizing, but I would suggest a large if not the largest percent will be down to sizing issues. New digital tools can be used to harness actual customer sizing data in the form of hardware scanners or even smart phone scanners, each offering vast improvements on traditional body scanning.

The same data could then be used in the same way that we use trend data, to dynamically manage our customer fitting and size range management. All of which could be pushed to PLM and 3D to improve the accuracy of design and fitting.

3D: creation to manufacturing

There are two primary paths for 3D. The first is to create and use digital assets (including avatars linked to your customers; body types and sizes), and styling details utilising 2D pattern shapes coming from your 2D CAD systems or styling, cutting the patterns whilst on your avatar in real-time. Then, bringing in your materials, trim and components from your digital platform(s) made up of the real and digital assets relative to the path and need. You can experiment with scaling across the size range, grading logos, resizing prints etc.

Once you are happy with your new creations, you can improve the quality of the images using creative blending tools and smart lighting, or even using backgrounds that help to place your new designs into a realistic virtual scene setting. Or even push directly to your ecommerce channel as a virtual-twin ready to sell or test demand.

Once you are happy with your 3D aesthetical design, you can break the details down into a 3-dimensional specification that includes colours, materials, trims, components, pattern pieces, seam types, seam allowances, lengths and shapes of seams, thread calculations linked to seam types and so on. This results in a 3D BOM and synthetic BOL.

2D automatic pattern & marker making

We have already touched on KYC (knowing your customer) sizing data. Today, using AI, we can extrapolate the sizing data in the form of a size chart and automatically build pattern pieces in the way that a tailor would create a made-to-measure set of patterns and a finished tailored garment.

Then taking the same pattern set along with the demand – be it for a single sized, personalised garment order or using an aggregated size curve to create the blended size order based on actual real-time demand curves – to automatically make the AI-driven cut order planning to define the number and designs of the markers or NC-cut paths. Again, using AI to drive the knife, resulting in a higher quality of cutting and cut path efficiency.


The data gathered during both the creative and development processing will be shared in real-time at various stages of the process, sending and receiving data from the PLM solution. This same data could be used to push to new on-demand processing – for example, digital printing directly onto garments or directly onto rolls of materials. Going forward this is a must for those businesses that require speed, efficiency and at the same time sustainability.

The synthetic BOM & BOL coming from 2D and 3D can now be used in PLM to create the total costing based on the order volume and size ratio.

And what about implementing machine learning in PLM? Always the big question is how much a style will cost. In the early stages of design and manufacturing we often use the term of a style following the 80-20 rule, referring to a standard product (like a ladies blouse or a men’s shirt) that will utilise around 80% of the styling and components that are found in a similar product type. In other words, around 20% of the new style will be the newness linked to new material compositions, styling, colours and construction methods, trims and the consumption of materials, manufacturing labour, size requirements etc. So, if this is the case, then isn’t it reasonable to question if PLM can use machine learning to automatically suggest 80% of a style’s basic requirement, auto-driven via basic information (product typespecialitygendermarket channels, collection, seasonality etc.). With the PLM solution automatically generating standard critical path, image views, size charts, measuring points, BOM, BOL, sample management, costing templates, RFQ’s, sourcing partners and so on.

Can PLM go beyond the central database or ‘single version of truth’? By building the use cases and systematically using data, can PLM be ‘intelligent PLM’? Unlike other systems, PLM has been very subjective; the user decides what to add into a style which could lead to delays and mistakes. But if it can also add an objective angle then, based on predictive analysis, right brain and left brain, can PLM solutions of the future help to develop beautiful products that are commercially successful?