From LED Lights to Electric Cars: The Experience Curve and How It Affects Us

In Austin, like many cities, the AM radio dial is eking out an existence based on sports talk, conservative talk radio, foreign language music (chicano and conjunto Tejano here), and the occasional oldies rock and roll. I'm a classic rock guy along with sports talk, NPR, and a little classical music when I need quiet and clear thinking; but occasionally I'll listen to several of the phalanx of AM conservative talk radio here in town in order to get their "take" on things.

The title theme above comes from commentary on the Sean Hannity program a year or two ago, during which he and a guest (his name long forgotten to me now) were just railing away on the U.S. government's decision to end production of incandescent light bulbs in place of the new CFL and LED technology. They went on and on about the abject stupidity of the decision, and were speculating that they might buy up all the remaining incandescent bulbs, inventory them until none were available, and sell them at a huge mark-up, yet still be much less expensive than the crazy new stuff.

Politics aside, neither had any knowledge of the experience curve. Most people don't. The seminal book on the matter, "Perspectives on Experience, was published by The Boston Consulting Group in 1968, and I have the good fortune to have a copy from 1973, designated "Assigned to J.K. George and property of Motorola Semiconductor Products Division." Hopefully good old Motorola, or what is left of it, won't come after it.

In 1973 Motorola Semiconductors still was a small but rapidly growing business. This book was influential in my career, and I just might have been the only person at Motorola to read it. Based on the mathematical models, Bud Broeker and I made $500 million decisions on memory chip industry prices, and fortunately never made an inaccurate one.

A new technology, one fundamentally disruptive to the established way of producing a product, is more expensive than the older method at first. Initially it is used by "early adopters," users who understand the value and features and are willing to pay for them. As volumes increase, the cumulative "learning" or understanding based on more and more experience drives efficiency such that overall cost of ownership is equal and then less. At some point, the general user base converts to the new product; the transition is rapid, and the older product becomes inefficient to manufacture and becomes obsolete. Strategically, a supplier of the new technology must be fully committed and well-positioned at this point. The conversion or transition process can be plotted in terms of market share. The curve is known as the "S Curve" since the initial penetration is slow (early adopters, higher price) and then undergoes a fast increase, and finally slows down as the market share approaches 100% and only late adopters and the "I'm  never gonna use that new-fangled thing" crowd holds back.

As a fundamental element of Experience Curve theory, the cost in constant dollars (this is critical) goes down between 20% and 30% every time the cumulative unit volume for the industry doubles. At first, the overall unit volume doubles quickly, in a half year or so, but later it takes longer. The ability to determine the actual unit to measure is key. While this is easy for an automobile tire or a TV set, it can be complex for the memory chip industry, where the number of actual "bits" in a single "chip" go up by a factor of four every year and a half. In this case, the proper "unit" is each bit of memory, not the physical number of "chips."

Our pundit, Mr. Hannity, obviously was harping on the heavy hand of government interference. Fair enough, but he did not portray a balanced picture—usually these radio talkers with their one-sided opinion machines never do. Hannity didn't mention the overall government policy to encourage transition from fragile and short-lasting energy-hogging light bulbs that emit lots of heat. This requires lots of air conditioning, which could drive electric utilities to add expensive capacity. He didn't explain that the new technology not only would come down sharply in cost but would save huge national commitments in energy generating plants with concomitant reductions in emitted air pollutants from the power plants. Apparently, he was unaware of the rapid (and inexorable) cost/price reductions of a new technology according to the Experience Curve.

At times, technology seems overwhelming as waves of new developments appear. Just as surely as bias ply tires (remember those?) were replaced by radial tires, as analog TVs were replaced by digital TVs, as cell phones transitioned from analog to digital to smart phones based on Internet connectivity, we can be certain that new forms of lighting, probably LED based but possibly other new forms (see organic LEDs), will become standard in our homes and offices. One of the biggest transitions in our lives, the conversion from petroleum-based vehicles to hybrid-electrical cars and then on to full-electrical automobiles will take place within fifty years. The initial movement to electrical powered cars is underway. Every one of the automobile companies is committing to it. Some day, our kids and grand kids will look back and wonder how we ever drove those old vehicles. Change is coming.


6 Responses

  1. Ben Bentzin
    The experience curve is real, dramatically driving down the cost of the new technologies as time passes and volume goes up. The experience curve, however, doesn't require or benefit from government mandates. No government intervention was required for Apple to develop touchscreen smart phones or Google to develop a self driving car, these technologies are moving up the experience curve because of private investments and consumer demand. A normal experience curve, without government intervention, would have driven down the cost of LED lighting to the point where low power consumption, low heat, and long life would have caused LEDs to naturally take over the lighting market. Artificial government manipulation of technological development is subject to lobbying from industries and interest groups that can diminish innovation. Government does not have the ability to forecast technological winners, witness the $500 million of taxpayer dollars lost in the Solyndra bankruptcy. Technological innovation and the experience curve will help ensure a bright future for our nation through investment, innovation, and responsiveness to market demands, not government mandates.
  2. Well stated, as usual, Ben. Your viewpoint always is studied and thoughtful. Clearly, the product transition in televisions was market-driven, with the caveat that at one point, the government decided on a cut-off date for analog television transmission. The industry had moved on, but the government/FCC needed the frequency spectrum for other uses, including mobile phones. Would you have let a few under-capitalized and obsolete licensees continue run analog programming "forever," delaying the roll-out of new CDMA technologies requiring additional spectrum? So far, any government influence or "guidance" in the vehicle area is in terms of aggressive mileage standards, and also some grants for electric car technology and possibly other areas. You and I have a fundamental difference in the role of government, as I agree with a moderately active government role. Clearly, the government should not pick winners and losers, but I don't buy a completely hands-off approach. That's a fair subject of debate and your position definitely is a valid one.
  3. Ben Bentzin
    Jim., I always look forward to and enjoy your thoughtful blog posts. My comments about limiting the role of government in selecting technological innovations are not a complete rejection of government. Airways are owned by the public and government should regulate the airways on behalf of citizens. Ditto with public roadways and waterways. For example, I support highway safety standards because trucks and autos are operated on public roadways. On the other hand, I am skeptical of government efforts to selectively invest in and promote electric car technologies. In the past five years the Obama administration has spent $5 billion+ on electric cars and I don't know that we have much to show for that investment beyond what the market would have done anyway. A healthy rise in the gas tax, funds dedicated to the development of our public roadway infrastructure, would do far more to spur development of electric cars then "targeted investments "or tax rebates. The airways are a public good and were given to broadcasters free of charge to spur the development of broadcast infrastructure. The move to digital broadcast was necessary to free up spectrum for other public uses, an appropriate application of government power on behalf of the citizenry to manage a public resource. Because of the influence of lobbyists, big businesses, and public interest groups, the role of government in selecting specific technologies should be severely limited. There is, nevertheless, an important role of government in administering public resources for the public good.
  4. Rob Brownstein
    Clearly this difference of opinion about whether and to what extent the government should get involved evokes lots of shades of gray. Without the government support for basic research, a lot of things that later become technologies would never be developed, or would develop much later. Private companies, with few exceptions, are reluctant to fund basic research. The ROI is simply not there. Governments do not need an ROI validation. Had the space program been subject to ROI validation, there's a good chance that we would not have, or not have had until much later, ICs and everything they fostered in turn. The space program also sped up research on materials that we all benefit from. I am not sorry to see the government stepping in to hasten the replacement of incandescent lighting with alternatives. Of course, it might have happened, later, but so long as incandescent pricing was so much lower, most people would have opted to keep buying them, much longer. Actually, I would like to see the government do more to encourage adoption of alternative energy technologies. Coal may be cheaper...but it is a lot dirtier. Whether you agree that green house gases are causing climate change or not, reducing them seems prudent.
  5. Comments from emails. The authors have been disguised. From PC: Thanks for sending me this. Very interesting and well written. From MK: Really enjoyed this perspective on the experience curve. From Dr. J: You "done gud agin." Posted it on my FB page. From LH: Very nice! Congrats on another interesting blog entry. From CK: Very cool Jim - thanks for alerting me. Could we reprint this in our company blog? From PS: Too bad there are so many books, so little time. That book sounds like an interesting read. Thanks for the informative articles - I look forward to your blogs. From ID withheld: Jim, your description of how the S-curve works is right on. Early adopters of a higher priced new technology do so because it is of a benefit to them. They can justify the higher price and then as others start to see the benefit, they start buying and the prices start to come down. This continues until the old technology is gone. Very good, right on. But what you are doing is using the LED light bulb to justify a Socialistic, Totalitarian and oppressive government to tell me how I should live my life. My brother, who is a socialist like you, has told me to my face that I live in too big of a house, that I drive too big of cars, and that I have too much money. If he and the likes of this central planning government (Chinese style) had their way, he would ban houses greater than 1200 sq ft, limit car hp to 125, and limit thermostats to min of 82 in the summer and 65 max in the winter, and confiscate most of my money. That is what out of control governments do. Maybe we would have been better to stay under King George (of England, not Bush) rule. Their GDP is still very strong while we are only one quarter from another recession. The light bulb edict is not about new technology; it is about a government that thinks it knows better than me and thinks it can decide how I should live my life. This is far from American principles of freedom, but is right out of the play book of Stalin, Castro, and Chavez. If you want to live in a socialistic country then go to Cuba. They have lots of rules. I want you and your kind to leave me the f… alone. I will pay my taxes, but you have no right to tell me how I should live my life. Of course there are reasonable limits on things, but this government and people like you and my brother are now way over the top. From DB: Enjoyed the travelogue as well as your blog on the Experience Curve. Wonder how many folks remember that UPS used battery operated delivery vans in New York City for many years. I recall seeing some in service in the 1950-1960 era. Of course they aren't the efficient batteries we have today but I appreciated the lack of smog! The only time AM radio is turned on in our household is when there is a weather emergency! From GD: Hi, Jim. Read your Blog on the Experience Curve. Good over-view. I lived thru a few cycles, with you, on semiconductor products. The semiconductor business was (and probably still is) a complicated business - not recommended for those who are faint of heart. From BW: Excellent. A wonderful read. Many thanks. From BK: Jim, I enjoyed reading your last BLOG on Learning Curve. It caused my mind to retrace some of the past. As I recall, 1973 was the year that I had the first opportunity to work with you. Those were tumultuous years but in retrospect lots of fun.
  6. The piece reminds me that in my molecular structure lab at USDA we recognized a principle called "Gresham's Law."* It was this: Say an existing piece of equipment does experiment A, B and C. A new piece of gear is bought to do the new experiment D, not possible on the old machine. If the new gear does experiment C 20% faster, Gresham's Law says that >80% of the time C will carried out on the new machine, time permitting. Recently, Barbara and I bought a Toyota Corolla Hybrid, -- 43 mpg driving to town (9 mi one way). The Subaru Outback, still a great farm car with very high road clearance and ability to pull a small trailer for hauling horse manure and the like, gets maybe 23 mpg for that trip. Applying Gresham's Law, 80% of our day-to-day trips are made in the Toyota. *When I searched for Gresham's law a few years ago there was no reference. It was a favorite saying of our group leader Tom Schatzki (MIT Physical Chemist) who had come from the old Shell Development research labs when they were moved from the SF Bay area to Houston. Now you can read the "official" economics definition of Gresham's Law on Wikipedia at . . "The law was named in 1858 by Henry Dunning Macleod, after Sir Thomas Gresham (1519–1579), who was an English financier during the Tudor dynasty." Maybe originally stated by Copernicus but there are even earlier versions, maybe going back to Arisitophanes in the 5th Century BC.

Leave a comment