The Game-Changing Technology of Generative Design
Generative design is touted as the killer app of the supply chain. As a result of well-modeled technology being fed large volumes of relevant, high quality data, it becomes possible to produce manufactured product designs that are beyond human capabilities. As supply chain leaders it is critical to understand the strategic value of algorithmic generative design and to determine if there is a use case for it in your product design and manufacturing processes. The power of the machine-based process to generate millions of possible design outcomes based on constraints and outputs has already created products and engineering components that appear to have been designed through organic evolutionary processes. The designs resemble complex natural organic forms and are practically impossible to manufacture by techniques other than 3D printing. Such was the case for UA when their engineers and designers went to work in the UA Lighthouse to take the HOVR shoe concept several generations further in its evolution.
Big Data + Generative Design +3D printing = A Totally New Kind of Product
The UA example boasts the number one asset to enhance the use of generative design algorithms in design and manufacturing: a large source of actual end-user performance data. The key to this story is the pool of sensor-based data generated by and freely submitted to UA for design analysis and algorithmic training. In some published examples, engineers place sensors on a modified existing product and gather data points during test runs. This sample data set is then fed into the generative design algorithm along with constraints to create design output options. Machine learning accuracy, however, depends on large volumes of data created under a wide variety of real-world conditions. This is exactly what HOVR and MapMyRun do for UA.
There are, of course, limitations to the use of generative design processes and some obvious reasons why it is not the blanket solution for design problems where aesthetics are more critical than performance. In the UA case, the ultimate generative design output was optimized for athletic performance, not aesthetics. The new shoe design was intended for athletes who are interested in, and willing to pay for, a performance edge from their equipment. This culture of performance over looks has many examples in the sport of running. It is not uncommon for runners to take knives to their running shoes to improve their fit, flexibility, and traction. Another example is the continued niche popularity of unfashionable running shorts which are perfect for the unrestricted movement needed to improve running performance but are horribly out of place in a restaurant or family gathering. As we have seen, it is all about best meeting the needs of your customer segment.
The Design Outcome: The 3D Printed UA ArchiTech Shoe
The UA team, armed with user-level running data, Autodesk generative design software, and 3D printing equipment, prototyped a new kind of futuristic running shoe, called the UA 3D ArchiTech. The shoe’s mid-sole looks like a complex web of flexible interconnecting rods and channels, unlike anything else UA had ever designed with human designers. The 3D ArchiTech will be outfitted with UA Record sensors, similar to the HOVR, to gather end-user feedback about the prototype’s performance. Imagine a development loop of printing, then making prototypes and measuring performance in the real world, iterating further on the growing body of design knowledge. Once a baseline design is established based on the large volumes of Big Data from users, how hard would it be for UA to establish a bespoke, custom printed ArchiTech pair for high- performance segment enthusiasts based on unique running attributes, habits, and form parameters? For customers willing to pay for this individualized performance product to gain a competitive edge, this may not be far beyond the realm of possibility. After-market orthotics, a helpful add-on for many runners may become obsolete as the shoes themselves can be manufactured for a customer of one, taking into considerations their unique physical needs.
What Does This Supply Chain System Look Like?
For one of our recent executive education programs, we created a one- page chart to illustrate the interconnections of the UA example which is shown in Figure 6.1. Some additional potential data points of social
Figure 6.1 UnderArmour ArchiTech Example (From a lecture by Kurz, 2020).
media reactions to the shoes as they are marketed were added to complete the picture. Commercially, the ArchiTech is still limited to prototypes as of this writing. At the prototype stage, a pair of ArchiTech shoes are prohibitively expensive to all but the most dedicated athletes. By sharing this example, the hope is that supply chain leaders will begin to think in a more integrated way across a user experience landscape when they re-imagine or design their DSCs. In one of our executive leadership programs, we had Craig Jones, UA’s Head of Supply Chain join us and we asked him for his reaction to the chart, first to see if we had the key components characterized correctly. To our relief Craig confirmed that it was basically correct for the purposes of describing the design and development strategy. The next key point that emerged in our discussion came from looking across all of the various departments, functions, data sources, technologies, talent, and feedback loops that are needed to bring this particular DSC to life. This picture, complex in some ways, and elegant in others, characterizes the leadership challenges supply chain executives face when seeking digital transformational performance outcomes. The relationships, influence, negotiations, interdependence, collaboration, and, ultimately, integration needed to bring the DSC to life is absolutely a leadership challenge.