To try to understand the mask shortages of 2020, you need to first try to understand global supply chains and the role that semiautomated digital platforms play in managing them. The problem is that those platforms are inherently hard for individual observers — and even their users — to comprehend, not only because of the complexity and scale of what they manage, but also because of how they were purposefully designed and developed.
“Software is eating the world” has become a familiar Silicon Valley rallying cry in the decade since venture capitalist and investor Marc Andreessen first coined it in a 2011 Wall Street Journal op-ed. Meant to encapsulate the idea that data-driven, software-based technologies will disrupt all existing industries and infrastructures to the extent that they become almost unrecognizable, there’s perhaps no better illustration of the concept than the automated transformation of global supply chains.
In a process started decades before Andreessen’s snappy slogan was put into print, management of the vast networks of factories, distribution centers, ports, ships, trucks, and retail stores has been consumed by sophisticated real-time, semiautomated data analysis and organization software. But like so many of Silicon Valley’s most beloved buzzwords, it’s a phrase that does more to oversimplify than to explain. Scratch below its shiny surface and you’ll find increasing levels of complexity, decision-making obscured by black box algorithms, and a largely hidden but very human army of labor.
“I grew up in Silicon Valley, but on the wrong side of the tracks,” Miriam Posner, an assistant professor in the Information Studies Department at UCLA, tells OneZero. “So it was always so fascinating to me that Silicon Valley was presented as this glossy, gleaming gold mine, when that wasn’t the reality that I saw at all.”
This interest became one of the main drivers for Posner’s research, borne in part out of wanting to teach students how to look beyond the facade to the globally distributed human workforces that keep the tech industry running. One idea Posner hit upon was to ask her students to take an electronic consumer device and trace all its components back to where they had been sourced or manufactured. She presumed this would be a relatively straightforward if laborious process.
“But in advance of the class, I tried to do it myself, and it was immediately obvious that it was impossible to trace all the components of any device,” Posner explains. “It was partly because the companies were keeping their suppliers close to their chests, but it also became clear to me that… they actually didn’t know themselves where their components were coming from.”
Posner eventually came to the conclusion that the only way she could get an idea of how things actually worked would be by learning the software platforms used to manage those supply chains herself. She decided she’d look into the platform developed by German company SAP, the world’s third-largest publicly traded software company and the closest thing supply chain management has to an industry standard. Platforms like SAP are used to coordinate the manufacturing and shipping of not just electronic devices, but pretty much everything — including masks and PPE. After poring over marketing materials and instruction manuals, Posner decided the only way to get a real feel for how SAP worked was to pay to take an online course in its use, which would allow her to get her hands on the software.
“[Everybody] is just sitting there waiting to get data from someone, so if that data doesn’t arrive, or it gets garbled, or whatever, there’s not much you can do.”
What she eventually saw was very far removed from some kind of omniscient “god view” of supply chains — it was quite the opposite, in fact. SAP by design atomizes the process into individual modules, with users able to see and control only the small parts of a given supply chain they’ve been trained on and given access to. The end result is a workflow that in many ways feels more like a production line than a management tool.
“Everybody works on a very different module,” Posner says, “and the data gets handed over from module to module as a package that then needs to be unpacked and then put to use by the next expert.”
SAP doesn’t give users any real sense of scale or indication of what might be happening at other points along the supply chain.
“It’s just not designed to,” Posner says. “When you’re working in a module, your goal is to squeeze all the latency out of your step of manufacturing and then hand that package over to the next person, who will do the same thing with their step in the manufacturing.”
For Posner, the real weaknesses of this production-line approach to decision-making have been exposed by the product shortages during the pandemic, something she’s watched by hanging out in industry spaces online. “The managers that I’ve encountered on these boards and webinars, they feel like something has happened to them, and they’re going to respond as best they can, but they don’t seem to feel like they have much control realistically,” she says. “Which makes sense, because [everybody] is just sitting there waiting to get data from someone, so if that data doesn’t arrive, or it gets garbled, or whatever, there’s not much you can do.”
And while SAP’s siloed, distributed approach to management might seem like an efficient solution to handling the incredibly complex processes its users deal with every day, Posner is keen to stress how fragile and risky it can be when it needs to deliver essential supplies in an emergency. “In order to get those items to the right places, you actually need someone to sit there and make decisions and coordinate where things are going to go,” she says. “And not in terms of how many people are willing to pay for them, but how many people need to have them. But there’s no one sitting in the middle of these supply chains dictating where PPE and vaccines, etc., need to be.”