Type in any combination of ‘robotics’, ‘smart factory’, or ‘IoT’ into Google, and you can spend hours blissfully reading about robots that teach themselves, smart machines that teach each other, AI, AR, IoT, and so on. Buy into the hype entirely, and the ubiquitous smart factory begins to seem like a veritable certainty – so much so that you enthusiastically hand over your hard-earned cash for a piece of the pie.
That’s exactly what I did right after Tesla unveiled its groundbreaking Model 3 in May 2016. Preorder for the so-called luxury electric car for the masses eventually grew to 500,000. It is now almost two years later and Tesla has delivered a paltry 3,000 cars, exactly none of which have rolled off a fully-automated production line.
So is techno-utopia just around the corner? Or did we all just give Tesla a $500,000,000 interest-free loan?
Gigafactory 1 OR How Elon Musk Spent His Summer Vacation
In Elon Musk’s vision of the future, every Tesla product will emerge fully-formed from a vertically integrated factory, where material delivery, manufacturing and assembly will be done without human intervention. Musk describes his smart factory of the future as a sort of “product” in itself. His proof-of-concept is Tesla’s new Gigafactory 1 in the Nevada desert, meant to be the most automated plant in the world. Surely Tesla should be on the bleeding edge of advanced automation. So why aren’t shiny new Model 3’s rolling off fully-automated assembly lines at breakneck speed?
Today Tesla is going through what Musk calls “production hell” as it struggles to ramp up the Model 3 and fulfill a mountain of orders. Just last week, Musk told reporters he’s so busy that he’s stopped showering and taken to sleeping on the factory floor, as his team continues to wrestle with numerous production and quality issues. On 16 April it was reported that Musk had ceased all production of Model 3s in order to fix bottlenecks at the factory. Several hedge funds are betting big that Tesla runs out of money before the issues are ironed out and many customers have begun to pull out (as of the writing of this article, pre-orders were down to about 460,000). What’s going on here?
From Model B to Model 3
Almost every industry is currently experiencing an explosion of SKUs. Just over a decade ago, the typical order for shoes was 100,000. Today it’s about 1,000. CPGs in particular are grappling with having to produce ever more variations of their products as we move towards a small batch world.
This presents manufacturers with a whole new set of challenges, as truly-automated production lines are inherently expensive and rigid. They require highly precise machinery adapted to very specific sets of tasks, and are slow and expensive to adapt to a new set of tasks. This means switching production from, say, minimalist running shoes one day to technical cross-trainers the next, is highly impractical.
In a small-batch world, competitive advantage in manufacturing can no longer be built on massive economies of scale
In some sense, this isn’t news. More than 100 years ago, Henry Ford showed the world that automation works when the products are simple and uniform, as with the simple Model T at his plant in Highland Park, AND that it quickly falls apart as soon you add complexity or variability, as with the more complex Model B at the plant at River Rouge. Toyota later rediscovered this insight after several failed experiments with automation in the late 1970s.
The failure of the highly-automated Lexus LS400 plant in Tahara, Japan, left a lasting impression on Toyota. Sales were below expectations and the plant was under-utilized. The key lesson was that capital costs are fixed and could not be adjusted to match demand. Since then, Toyota has actually reduced automation rather than accelerating it. Today, Toyota’s principles of production equipment are “simple, slim, and flexible” to work in harmony with people. The company prides itself on only building to actual demand.
As Jeffrey Liker points out in Tesla vs. TPS: Seeking the Soul in the New Machine, Tesla might be missing exactly what’s made Toyota so successful. Toyota manufacturing is built on a “living system approach” that integrates advanced robotics with people trained on TPS, resulting in a production system that continuously improves itself and can quickly adapt to demand and other variables.
The Future of Industrial Work
So will the next chapter of industrial work be defined by near-complete automation and mass unemployment, or will humans continue to play a central role in industrial work?
The history of automation is one of huge successes as well as notable failures. This is a trend that’s likely to continue. While technological advances underlying Industry 4.0 seem to bring lights-out manufacturing ever closer, other parallel trends such as SKU explosion are creating a “long-tail” for consumer goods. This added complexity and variability makes the task of automation exponentially more challenging. This may help explain some of the issues Tesla is facing on the Model 3 production line.
As human workers and machines must increasingly work together in symbiosis, what’s clear is that the nature of industrial work is in flux. The rules for how a modern industrial firm operates must also change. Simply automating away as many human jobs as possible is clearly not the answer, and can lead to expensive missteps. In a small-batch world, competitive advantage in manufacturing can no longer be built on massive economies of scale. It must instead be based on thoughtful integration of smart machines with human worker in collaborative digital work processes, for a truly “connected” future of industrial work.
There are some glimmers of hope. On 13 April, Elon Musk Tweeted this:
Appropriately, the factory Musk currently calls home in Fremont, California (and where he likely penned this Tweet) once housed NUMMI – the partnership between Toyota and GM aimed at bringing human-centric TPS practices to US auto manufacturing.
ABOUT THE AUTHOR
Edward Gorenshteyn is Director of Product Marketing at Parsable, the mobile collaboration and workflow platform. Parsable is an official Partner at Manucore’s Manufacturing Leadership Forum USA in Atlanta, May 15-16.