Elon Musk and “First Principles” Thinking
Improving the factory performance by an order of magnitude
Disclaimer… I’m a Tesla fan… of the car (mine is #S00067) and the thinkers behind it and the original inventor as well.
Where Elon Musk really got my attention was when he mentioned that he was beginning to rethink the factory and saw the possibility of improving efficiency by orders of magnitude.
“We realized that the true problem, the true difficulty, and where the greatest potential is — is building the machine that makes the machine. In other words, it’s building the factory. I’m really thinking of the factory like a product.”
The factory capacity is “volume x density x velocity … and we use maybe 2 or 3%.”
(Old video link is broken — new link: https://www.youtube.com/watch?v=f9uveu-c5us)
You know, you may be right… it’s possible to increase productivity 10x
Here are some basic principles I’ve learned working in manufacturing and studying how to solve problems. While no one I know has gone as far as Elon Musk’s “volume x density x velocity” reduction to first principles, many have started in that direction and identified some of the principles a step or two away that might be useful to think about.
There exists in history a few rare examples of improving a whole factory an order of magnitude. This proves that it’s possible, and that might help others to better understand the potential of Elon’s vision for improving the efficiency of production in his facility.
An example can be found within Ford Motor Company, not once, but twice. It is the second example that people are most familiar with… the implementation of a moving assemble line. It was the first improvement cycle that created the foundation for the second.
About 110 years ago, the board of the Ford Motor Company thought the production could be improved, so in the fall of 1906 they brought in Walter Flanders to help improve production. They had not quite produced 1600 cars the year before. The first thing Flanders did, over the objections of the accountant, was to rearrange the whole production floor. He went on to work reorganizing the supply chain and developing the distribution network. He closed the 1906 year with sales of 8,729 cars. The next year saw production increase to 14,887 cars. This was done with fewer people, in the same shops. In 18 months, the productivity of cars per person per year jumped from 2.28 to 25.89.[1]
Walter Flanders had achieved an order of magnitude improvement, in an extremely short period of time. More importantly, he taught a handful of men the ‘flow principles’. With their successes, they were infected with the idea of unlimited possibilities to improve the manufacturing and the assembly processes. Henry Ford realized that this could lower the costs and make the car more affordable, so he created an environment that supported, even pushed, experimentation and innovation. The development of parts production continued faster than the development of the assembly processes. Eventually they would use the same principles and develop their moving assembly lines.
Their learning curve was based on iterative development; where each expansion allowed experimentation with new ideas that included building construction (layout) as well as how all the processes interconnected. The most efficient Ford factory would be the last building in the Highland Park complex, the ‘new shop’. Similar to the 4 story ‘old shop’, this one was 6 stories with a railroad spur into the center of the building. Materials were off-loaded on their designated floor and as they flowed outwards and downwards the parts became completed into sub-assemblies to be fed to the three assembly lines on the first floor. Everything was balanced to reduce transport as much as possible. Even the conveyor lines were carefully thought out. They were long enough for an item to cool after coming from a furnace, or the paint to dry before the next operation, or just enough buffer for the processes that were not yet stable.[2] The railroad cars of parts were sequenced in a yard outside for just in time delivery during the production hours.
Courtesy of Albert Kahn Architects.
The thinking behind the ‘flow principles’ did not materialize overnight or out of thin air, factory rationalization preceded it.
Factory rationalization was partly driven by the fascination with inventions in the 1700’s and 1800’s to solve the challenges of manual work. Work that had previously been purely manual tasks, now had some portion of the work done with the aid of a mechanical device; this eventually spilled over into rationalizing factory designs. The principles of increasing productivity had begun to be recognized, not just as isolated inventions, but could be focused for improvement efforts. A look back in history will show that the improvements served different purposes in different industries. Across various industries, you began to find advocates talking about the whole factory as a living machine; shifting from improving with individual machines toward systems thinking as they focused on synchronizing all processes. They addressed their system constraints with focused improvements. Our generation will have machines of a higher order; such as robotics with machine learning (AI), where the machine will begin to learn how to improve its own performance.
Of the three principle factors, velocity is the primary principle constraining factor for factory output. Output is not the average of all the process speeds; it can be no faster than the system’s slowest and most inefficient process step. To improve the system, the whole factory, you address your weakest process functions and synchronize the system.
Rationalization — Solving the Constraints
The challenges, obstructions, snags, barriers, problems… whatever you want to call them, fall into four general categories, or orders of improvement.
1st order improvements were to multiply the productivity of the skilled craftsman.
The focus at this level is often seen where there was a talent shortage of skilled craftsmen and high demand for the products that they produced. Here they would improve the tools or develop new machines to increase the output capacity of a craftsman.
2nd order improvements were to reduce the skill requirements needed.
This is a continuation of the 1st order improvements; to reduce the dependency on scarce skill sets. By embedding some of the expert’s skill in the machine, simplification of skill requirements addressed availability of skilled labor.
With the advent of computer systems, we have taken this even further by embedding feedback loops or knowledge bases into equipment. An example is the anti-lock braking system on automobiles. The average driver can stop as fast or faster than an expert driver could before the technology was developed.
3rd order improvements were to reduce variation (leading to interchangeable parts).
The wooden clock industry pioneered improvement of processes and tools to create more consistent parts. Reducing the need to “fit” individual parts into an assembly. This increased the productivity of the skilled labor as well as reducing the skill level needed to assemble the clocks.
Lean companies seek to develop standard work which reduces the human factor variable; making problem solving and process improvement easier.
4th order improvements addressed transportation between processes.
Part of Henry Ford’s drive to improve the production processes was his observation that the most common workman in his shop was the man with the hand truck to move parts between processes.
Today, a more common perspective would be approaching constraints from a ‘Cost-Quality-Delivery’ (CQD) perspective. Every process step has a value in each CQD vector, which means the system can have multiple constraints (one constraint in each vector).
Rationalization may solve some productivity issues, but it does not build a production system. The Flow Principles offer a foundation to establish a synchronized production system.
Basic Principles of Flow
The basic principles of flow can be described in four steps; with the first three primarily happening within the four walls of the factory and the fourth driven mostly by sales and marketing.
1. Put Processes in Sequence
This is more than just moving machines or workstations around; it may require rethinking of how you perform the process. For example, typical stamping operations are noisy and can have cycle times of a second or less. This creates a synchronization problem with adjacent processes. Plating operations or heat treatment operations are often isolated departments. Innovative solutions rescale the process to match system needs.
2. Synchronize Processes
When sequential processes are grouped, they rarely are produced at the same cycle times. Early synchronization efforts started with staffing each process with sufficient men and machines to balance outputs.
Synchronizing starts with pairing two sequential operations when you achieve similar cycle times. As you begin to closely couple more sequential processes, cells emerge (natural limits are 3–5 processes between buffer stocks).
Grouping sequential cells or a series of processes into a production lines is the next level of synchronization. Flow begins to happen when the whole system is synchronized to produce at the same rhythm (the word Takt is German for beat or rhythm).
3. Balance Work Content
One of the hurdles that becomes obvious when sequential processes are placed close together is the cycle time differences. Solutions can include speeding up the slower processes until they are the same pace as the fastest process (or slowing down the faster processes to match). This may be a solution to address a few of the process, but not all of them.
4. Balance Work Pace
Walter Flanders introduced this practice to Ford’s plants in 1907 where he defined stable production rates on a monthly basis by setting targets from the distribution network where about 80% of the cars produced were made to order. This allowed smoothing the orders to their suppliers as well. Heijunka, the Japanese term for leveling production is best applied with the coordination of the sales and marketing organization, not just inside the factory.
By 1920 the basic Principles of Flow had been established and factories around the world were being reorganized for flow as people realized that the effective implementation of the flow principles would yield a competitive enterprise advantage. The principles sound simple, yet implementation is much more difficult that one would first anticipate.
Now all you now need is a structured problem solving process to start building your production system.
Building your Production System
One of the biggest hurdles to implementing flow is learning to see the disruptions to flow, otherwise known as a “problem”. Once you have identified a symptom of flow disruption, the problem can be better understood for choosing the next steps. The second hurdle is having a structured problem solving approach to resolving the disruptions to flow.
It is one thing to improve a single constraint in a system an order of magnitude, but a whole factory will be extremely problematic.
Taiichi Ohno’s (Toyota) breakthrough came when he incorporated the structured thinking patterns of Training Within Industry’s (TWI) Job Instruction, Job Relations and Job Methods programs with his attempts to create flow. The TWI skills created a thinking structure that everyone could follow to solve problems; it even including coaching of the leaders. Ford’s greatest asset was not the moving assembly line, but the people, the thinkers that invented it. His competitors had assembly lines within 6 months, but not the thinkers who developed it.
With the thinking patterns known, the next difficulty is how to develop the awareness of people to see the disruptions to flow. To make it easier to “Learn to See” the symptoms of disruptions to flow, Ohno created the list of ‘Seven Wastes’ and applied the thinking patterns to solve the disruption problems identified. To resolve the disruption, you first understand why it exists, then develop solutions to eliminate or contain the issue.
Once you have “learned to see” and discovered lots of opportunities, now you need to decide which ones to take advantage on or implement. There is an optimal sequence of improvement; work on the constraints in each vector. Not just the physical layouts that interconnect, but the information flows as well, every interconnected node, tangible or intangible. As you practice, more opportunities will be discovered.
Elon Musk observed that an engineer, on thinking about first principles for factory design, will see lots of opportunities. Imagine what would happen if all your managers, supervisors, team leaders… every employee understood that there was a structured sequence to problem solving to improve the processes, not just on the shop floor, but in the support functions as well. Ford’s advantage wasn’t the moving assembly line, but the inventive people that did continuous improvement. Creating an environment to grow people is a competitive advantage.
Velocity — the Principal Constraint for Tesla Motors
Achieving a stable velocity depends on flow not being disrupted…
To begin to meet the demands for the Model S, Model X and now the Model 3, Tesla Motors will need to ramp up to producing 500,000 cars per year in the near term. This will have the exit velocity exceeding 2 cars a minute, or approaching a 20 second cycle time.
This poses another problem… transport time erodes the ratio of value adding time in this 20–30 second window. The counter-intuitive solution is to slow down the cycle times to optimize the value adding time over the other four process functions. When you design computers, this is parallel processing… in a factory, you have multiple lines.
Further out, there’s the transport logistics of a complete car. Henry Ford had a similar problem as the demand for his Model T accelerated. They could ship only three fully assembled cars per railroad car, but 27–28 knocked down kits would fit in a railroad car. These were assembled at local factories with the larger body parts supplied by local manufacturers.
Connecting the Dots
Visionary leaders inspire others to join and make that vision a reality. This requires that the vision is one that is also in demand by others outside the organization and that the capacity to deliver the product or service is developed.
Elon Musk has created a vision for accelerating the advent of sustainable energy[3], and hundreds of thousands eagerly await the Model 3 arrival. The missing link is to see the vision for the factory of the future to its completion.
Elon Musk set a good example for other CEO’s and anyone in leadership a position in manufacturing where he talked about (2016 Shareholders meeting) spending all of his time of the production floor.
The time spent on the factory floor leads to observations, which lead to breakthrough thinking. It’s amazing what you discover when you go to where the work is being done, observe, question, and get the facts; then you can begin to see the real problems.
What can you do to help the people in your company be more productive? If you can improve efficiency of manufacturing by an order of magnitude, then there is no need to chase cheap labor and export jobs.
The Tesla factory puzzle is one that would very challenging and interesting to work on. I’m happy to see that Elon has concluded that substantial improvements are possible inside his factory and factories in general. I’m excited about the possibilities of innovation in factory design as seen in Elon’s other efforts.
“As to methods there may be a million and then some, but principles are few. The man who grasps principles can successfully select his own methods. The man who tries methods, ignoring principles, is sure to have trouble.”
– Harrington Emerson
If you found this post useful, please hit the little ♥ to spread the word and share with others.
[1] 1905–700 people produced 1599 cars = 2.28 cars per person per year 1907–575 people produced 14,887 cars = 25.89 cars per person per year
(100 years later, the leading companies produce about 80 cars per person per year… similar to Ford’s production rate in 1917.)
[2] https://www.amazon.com/Methods-Horace-Lucien-Arnold-Faurote/dp/1421253658 — reprint of a series of articles written as Ford’s men were developing the assembly line.
[3] https://www.tesla.com/blog/master-plan-part-deux — Tesla Master Plan: Create a low volume car, which would necessarily be expensive; Use that money to develop a medium volume car at a lower price; Use that money to create an affordable, high volume car; Provide solar power.
Master Plan part 2: Create stunning solar roofs with seamlessly integrated battery storage; Expand the electric vehicle product line to address all major segments; Develop a self-driving capability that is 10X safer than manual via massive fleet learning; Enable your car to make money for you when you aren’t using it.
If you found this post useful, please hit the little ♥ to spread the word and share with others.