Hitting the Wall—Part 1
By Alberto Moel, Vice President Strategy and Partnerships, Veo Robotics
Over the last few blog posts, we have been talking about a structural slowdown in manufacturing productivity and the reasons behind it. Figure 1 (reproduced from our previous discussions) shows data from the US Bureau of Labor Statistics (BLS) on YoY changes in TFP for the US manufacturing sector.1
As we can see, TFP in the US manufacturing sector grew by an average of 2.0% per year from 1992 to 2004, but from 2004 through 2016, manufacturing TFP declined by an average of 0.3% per year.
Figure 1. US Total Factor Productivity YoY Percent Change 1993-2016.
The drivers of this slowdown in TFP growth have been widely studied and attributed by many fine economists to the “one-off” semiconductor and IT revolution of the late 1980s and early 1990s driven by (the now-waning) Moore’s Law.2 Others argue that outsourcing or offshoring,3 reduced investment in R&D,4 or declines in business dynamism5 have led to reduced manufacturing innovation and weaker productivity growth.
Our take is that one big contributor to the slowdown is the way manufacturing is currently practiced, where innovations are incremental and human-machine collaboration is limited. As we have pointed out on multiple occasions, poor human-machine collaboration can be suboptimal and not sufficiently responsive to the long-term trends of shorter product life cycles and increasing product diversity. In some respects, modern assembly lines still resemble those from 100 years ago (Figure 2).
Figure 2. Ford Model T and a modern automotive final assembly line, 100 years apart.
In other words, we believe the historical inflexibility of manufacturing is partially to blame for today’s weak productivity growth. The current incremental model of manufacturing productivity improvement is not enough to generate value and long-term growth.
Productivity in manufacturing has hit a wall, and jumping over (or crashing through) this wall will require novel approaches to introducing flexibility in factories and workshops. Of course, our favorite novel approach is the Veo FreeMove system, which we believe will bring a material improvement in manufacturing productivity.
But the objective of today’s post is not to dwell any further on manufacturing productivity and how we can improve on it. 6 Instead, let’s do a bit of lateral thinking about how other technologies have jumped over or crashed through their own walls of stagnant productivity growth to see if we can glean some insights that can be applied to manufacturing.
Overcoming the wall of incrementalism
There are plenty of examples 7 of technologies that plodded along for decades (sometimes even for centuries) with only incremental improvements in some aspect of their efficiency or productivity, and then, one day, presto! an order of magnitude improvement (sometimes many orders of magnitude) pops up and revolutionizes the technology forever.
Consider, for example, the time it takes to transmit a 140-character message across the Atlantic, which is 1,900 miles wide. Well, before the first underwater transatlantic telegraph cable in 1858, the message (whether 140 characters or an entire encyclopedia) could only travel across the Atlantic via ship.
Ship speeds did not really improve much over the centuries, whether wind- or steam-driven. From the time of the Vikings until the first steamships in the 1840s, messages could not be delivered at speeds greater than 6 knots (or about 7 mph). Even steamships and their improved technology in the 1850s never exceeded 14 knots (16 mph).
And then, boom! in August 1858, the first 617-character telegraph message was sent through a 1,900-mile underwater cable across the Atlantic, taking about 16 hours to transmit. 8 With some simple number crunching we can convert that to an equivalent speed of 140 characters at 450 knots (517 mph). That’s more than a 30x improvement in one fell swoop. From there it was off to the races—through serious improvements in transmission technology, by 1928, a telegraph cable could handle about 400 words per minute, meaning a 140-character message would take just 21 seconds, effectively traveling at about 325,000 mph.
How about cargo? One cannot send cargo via an underwater telegraph wire, so before the (boom!) advent of the airplane, cargo had to travel by ship. To be fair, in 1919, when Alcock and Brown flew a plane non-stop across the Atlantic for the first time in 14.5 hours, ship speeds had improved to 26 knots (30 mph). Alcock and Brown’s equivalent speed was about 100 knots, not quite a 4x improvement.
Then came another steep improvement curve, culminating in supersonic flight. For example, the SR-71 Blackbird, which was introduced in 1964, could fly at 1,500 knots. That’s a 15x improvement over the original transatlantic non-stop flight in just 45 years. 9
Across centuries, cargo speeds across the Atlantic quadrupled from about 6 knots to 26 knots, and then in one eventful day in 1919, they crashed through the wall and quadrupled again.
The time for breaking the wall in manufacturing is now
The two examples above imply that those productivity explosions were independent and serendipitous. But in reality, the alternate enabling technologies (i.e., the telegraph and the airplane) had been decades (or even centuries) in the making. Only when applied to a specific metric (i.e., message and cargo speeds) could we see the multiplier effect.
Additionally, many new technologies are only productive in conjunction with new approaches. Though technological innovation may make the first crack in the wall, the way in which those new technologies are applied also needs to change in order for that crack to grow and eventually cause the whole wall to come crumbling down.
The manufacturing industry has been talking about introducing flexibility into factories for years, and progress has been incremental for just as long (or even longer). So why do we at Veo think we’ve reached the tipping point? How do we know we’re on the cusp of a technological transformation?
New computing architectures, sensing technologies, algorithms, and software design processes are now moving onto the factory floor. The Industrial Internet of Things (IIoT) and Industry 4.0 are becoming reality. And advanced intelligent software and hardware that makes industrial machinery aware of human presence and responsive to it, enabling true collaboration between humans and machines (i.e., Veo FreeMove) is now close to commercialization.
The methodical incorporation of these innovative technologies will make factories more flexible and subsequently change the way they operate. Very soon, flexible factories will become an industry norm, and when that happens, we can expect to see dramatic productivity gains.
In the next post we will get a bit more serious and present three examples that are very relevant to the approach Veo is taking to revolutionize manufacturing, so stay tuned!
1 Michael Brill, Brian Chansky, and Jennifer Kim, "Multifactor productivity slowdown in U.S. manufacturing," Monthly Labor Review, U.S. Bureau of Labor Statistics, July 2018.
2 See, for example, David Byrne, Stephen Oliner, and Daniel Sichel, “Is the information technology revolution over?” International Productivity Monitor, no. 25, Spring 2013, pp. 20–36.
3 Matthew Dey, Susan N. Houseman, and Anne E. Polivka, “Manufacturers’ outsourcing to staffing services,” ILR Review, vol. 65, no. 3, 2012, pp. 533–59.
4 Roberto Cardarelli and Lusine Lusinyan, “U.S. total factor productivity slowdown: evidence from the U.S. states,” IMF Working Paper, May 2015.
5 Ryan A. Decker, John Haltiwanger, Ron S. Jarmin, and Javier Miranda, “Declining dynamism, allocative efficiency, and the productivity slowdown,” American Economic Review, vol. 107, no. 5, May 2017, pp. 322–326.
6 With the Veo FreeMove system, obviously.
7 These two examples and the underlying data come from AI Impacts.
8 These were the first test messages in early August, preparing for the “official” test message which was a message from Queen Victoria to US President James Buchanan. That message was 509 characters long and took 17 hours and 40 minutes to transmit.
9 Although, we should finesse the fact that the SR-71 didn’t carry any cargo besides the pilot and a bunch of spy cameras.