How to create a market instead of fighting for share

1 The bottom line

Create a market by serving people no rival counts as customers, then teach them why to buy. Where there are no customers there are no competitors, and the founder who educates the market owns it first.

2 Key takeaways

• Go where customers do not exist yet. Rivals shape themselves for a market only after it appears.
  • Jensen Huang, Nvidia46:46 says Nvidia spent about a decade in zero billion dollar markets, so few competitors were shaped to compete when they emerged.
  • Bob Noyce, Intel38:28 skipped Fairchild's market and built memory chips to replace ceramic ringlets.
• Build for the many that incumbents ignore. Products built for a narrow few keep demand moderate.
  • Soichiro Honda, Honda34:30 redesigned for the man in the street, his wife and his kids. The Super Cub has sold over 100 million units.
  • Frank Robinson, Coca-Cola38:20 moved Coke from a brain tonic for the few to a refreshment for the masses.
• Size the market from structural shifts and real money pools, not from today's customers. Forecasts built on the present undercount a new category.
  • Bill Gates, Microsoft25:46: MITS forecast about 800 Altairs a year and sold thousands in a few months.
  • Jensen Huang, Nvidia1:32:37 replaced a claim of 1% of $100 trillion with $1 trillion of installed data center assets and $250 billion of yearly spend.
• Teach the market, because buyers cannot picture what does not exist. Education is part of the product, and whoever teaches first owns the category.
  • Jensen Huang, Nvidia50:15 gave over 100 talks in a year and wrote a CUDA textbook that sold tens of thousands of copies.
  • Akio Morita, Sony43:45 often held a new market alone for a year or more.
• Remove whatever stops a first buyer from saying yes. A new market needs supply, channels and legal room before demand can show up.
  • Irwin Jacobs and Andrew Viterbi, Qualcomm1:27:10 formed joint ventures for base stations and handsets so carriers could adopt CDMA.
  • Soichiro Honda, Honda29:05 wrote to 18,000 bicycle shops and ended up with 5,000 dealers.

Where there are no customers, there are also no competitors.

Jensen Huang, Nvidia

2.1 Enter where rivals do not look

Rivals follow existing demand, so the open ground is where demand is unproven or dismissed.

• Pick the area incumbents call too small or too backward Jensen Huang, Nvidia40:45
  • Tell the company to do what others cannot, such as building robotics where no customers exist yet. The work also attracts the best people Jensen Huang, Nvidia40:45.
  • Noyce and Moore did not compete with Fairchild. They moved into data storage, where computers still used ceramic ringlets, and had the 1103 memory chip within two years Bob Noyce, Intel38:28.
  • Intel had under $3,000 in sales in year one and $66 million by year five Bob Noyce, Intel38:28.
  • Apollo astronauts told Congress the market was too small for hundreds of launches, and incumbents would not spend to test it. SpaceX did, and Starlink now has over 9,000 satellites Elon Musk, SpaceX38:32.
  • Bezos built only what could not exist offline. A bookstore with 3 million titles was impossible in physical form, while Barnes & Noble and Borders stocked about 80,000 Jeff Bezos, Amazon48:04.
• Stay in the zero billion dollar market long enough for it to emerge Jensen Huang, Nvidia46:46
  • Position where a need has not yet emerged and keep going when experts object. A CTO said cars cannot tolerate the blue screen of death, but Jensen believed cars would become software defined Jensen Huang, Nvidia46:46.
  • Stand close to future opportunities so you can make a diving catch when one drops Jensen Huang, Nvidia36:45.
  • Omniverse is a current example, with something like 40 customers Jensen Huang, Nvidia46:46.
  • GeForce Now, built to stream games from servers, became Nvidia's first data center product. The company had learned data center building long before AI arrived Jensen Huang, Nvidia36:45.
• Choose a field you can dominate over a crowded one Walt Disney, The Walt Disney Company22:05
  • Disney saw that newspaper cartooning was crowded and animation was barely two decades old. He built his own curriculum from the one library book on animation Walt Disney, The Walt Disney Company22:05.
  • Chang asked what company would fit Taiwan's one strength and avoid its weaknesses. He picked wafer manufacturing Morris Chang, TSMC46:45.
  • Every major chipmaker made its own chips, and Jerry Sanders of AMD said real men have fabs. The idea looked bad at the time Morris Chang, TSMC46:45.
  • Walton found that a 13,000 square foot store in a town of 2,000 could do a million dollars in sales Sam Walton, Walmart18:55.
• Serve the customers a missing supplier is blocking Morris Chang, TSMC59:20
  • Chang saw chip designers at TI and General Instrument who wanted to start companies but could not raise money for a fab. He reasoned a foundry would let them start, with TSMC as their platform Morris Chang, TSMC59:20.
  • Qualcomm, Broadcom, Marvell and Nvidia grew up on this model. Nvidia raised only $20 million and never built a fab Morris Chang, TSMC59:20.

2.2 Aim at the many and name the real alternative

Markets stay small when the product is built for a narrow few, so redefine who the customer is and what they would otherwise do.

• Redesign the product for ordinary people, not enthusiasts Soichiro Honda, Honda34:30
  • The Dream was heavy, expensive and built for enthusiasts, so demand was moderate. Honda rebuilt around a machine that could substitute for the bicycle Soichiro Honda, Honda34:30.
  • Jobs judged that for every Homebrew hobbyist there were a thousand customers who wanted to buy a finished computer like an appliance. He ignored the club's lukewarm reaction Steve Jobs, Apple13:54.
  • The Super Cub touched off explosive demand and has sold over 100 million units in 60 years Soichiro Honda, Honda34:30.
  • A shop owner approached Jobs the same day, and a 50-computer order followed the next day Steve Jobs, Apple13:54.
• Change what the product is for Frank Robinson, Coca-Cola38:20
  • Robinson saw Coke's ads reached only people who needed a brain tonic. He shifted to "delicious and refreshing" and tied the brand to positive things, not problems Frank Robinson, Coca-Cola38:20.
  • Schultz realized Starbucks sold the experience of the place, with coffee as the conduit. He pitched community and the third place Howard Schultz, Starbucks29:54.
  • Wilsdorf saw that pocket watches were hidden and passed down, so the market stayed tiny. A wristwatch would be seen, worn harder, replaced sooner and bought in several Hans Wilsdorf, Rolex13:46.
  • Coke was in a soda fountain in every state and territory by 1895 Frank Robinson, Coca-Cola38:20.
  • A test coffee bar took the Fourth and Spring store from 200 to 300 bean customers a day to 500 in a week Howard Schultz, Starbucks29:54.
  • Wilsdorf's wristwatch orders soon ran into the tens of thousands Hans Wilsdorf, Rolex13:46.
• Define your competitor by what the customer would otherwise do Bill Gates, Microsoft1:03:57
  • Gates pitched that in-house software would cost X dollars, and Microsoft charged less than half of X for something more powerful Bill Gates, Microsoft1:03:57.
  • Kelleher told a shareholder Southwest competed with driving, not other airlines. He held fares at $15 when Braniff charged $62 Herb Kelleher, Southwest Airlines28:20.
  • Software was not capital intensive, so Microsoft made cash from year one even though many customers went bankrupt Bill Gates, Microsoft1:03:57.
• Celebrate the activity itself to widen the pool of buyers Phil Knight, Nike30:50
  • Bowerman wrote Jogging when ordinary people did not run. His view was that everyone with a body is an athlete, so the market was far bigger than elite Olympians Phil Knight, Nike30:50.
  • The book sold millions of copies. By 1970 Blue Ribbon had over $500,000 in annual sales, and 70% of US runners wore Onitsuka shoes Phil Knight and Bill Bowerman, Nike50:50.

2.3 Size the market you cannot yet see

Forecasts from today's customers undercount, so size from structural shifts, real money pools and cheap experiments.

• Assume a new product's market is larger than anyone forecasts, and test it cheaply Bill Gates, Microsoft25:46
  • MITS forecast about 800 Altairs a year, then sold thousands in a few months. Hobbyists drove hundreds or thousands of miles to Albuquerque Bill Gates, Microsoft25:46.
  • Southwest bet that low fares and good service would create new travelers. It tested the bet by flying into the Rio Grande Valley cheaper than every rival Herb Kelleher, Southwest Airlines18:54.
  • When a plane had to move for weekend maintenance, Southwest sold seats at about a third of the $26 fare. It was soon turning people away, which revealed a price-sensitive leisure segment Herb Kelleher, Southwest Airlines21:05.
  • Rio Grande passengers rose from 123,000 to 325,000 in under 11 months, a gain bigger than the whole prior market Herb Kelleher, Southwest Airlines18:54.
  • Kelleher would not fire people whose experiments cost $400,000 or $500,000 Herb Kelleher, Southwest Airlines18:54.
• Size the market from a concrete pool of money, not a top-down share Jensen Huang, Nvidia1:32:37
  • The 2021 claim of 1% of $100 trillion was blunt. In 2023 Jensen pointed to $1 trillion of installed data center assets, with $250 billion spent each year to upgrade and expand them Jensen Huang, Nvidia1:32:37.
  • Remember that the market, not the product, caps the company. Jensen was told no chip company could be very large. Selling the work a system does grew the opportunity perhaps a thousand times Jensen Huang, Nvidia1:25:16.
  • Nvidia's annualized revenue of about $40 billion is roughly 15 to 18% of that $250 billion yearly spend Jensen Huang, Nvidia1:32:37.
• Read the structural wave under the market Joe Coulombe, Trader Joe's47:30
  • Coulombe saw college attendance rise from 2% to 60% of graduates and the 747 make travel affordable. He built the store, name and shirts for the educated, curious, not-rich customer Joe Coulombe, Trader Joe's47:30.
  • Kamprad watched Swedish farms empty and young people move into apartments with no inherited furniture. He swapped rugged rural pieces for simple, light designs Ingvar Kamprad, IKEA53:00.
  • The 747 cut the cost of travel to Europe by 50% at once and by a factor of 15 within 10 years Joe Coulombe, Trader Joe's47:30.
  • IKEA's sales rose from 6 million krona in 1955 to 40 million by 1961 Ingvar Kamprad, IKEA53:00. About 50,000 farms closed in the 1950s Ingvar Kamprad, IKEA42:59.
  • Bezos read that web traffic grew 2300x in a year, and it jolted him into picking a first category Jeff Bezos, Amazon50:43.
• Watch for shrinking demand and odd customer signals Ray Kroc, McDonald's18:29
  • After 15 years of selling the multimixer, Kroc saw customers cutting back on milkshakes. He concluded he was in a dwindling business and had to find a new product Ray Kroc, McDonald's18:29.
  • Buyers kept asking for a mixer like the McDonald brothers had. Kroc checked and found eight multimixers making 40 shakes at once, with lines of cars Ray Kroc, McDonald's0:59.
  • He flew out at age 52 and signed with the brothers in 1954 Ray Kroc, McDonald's18:29.

2.4 Educate the market

Buyers cannot picture a product that does not exist, so teaching them is part of the product.

• Teach buyers why the product exists, and treat it as your job Bob Noyce, Intel50:08
  • Noyce, Moore and Grove ran a traveling road show for the microprocessor, which confused buyers and which many engineers dismissed as a gimmick Bob Noyce, Intel50:08.
  • Nvidia offered schools CUDA-capable machines if they committed to teaching a class. Its chief scientist gave over 100 talks and wrote a textbook because none existed Jensen Huang, Nvidia50:15.
  • Morita refined Sony's own thinking about a product and its use, then educated buyers. The public does not know what is possible, but the maker does Akio Morita, Sony43:45.
  • Intel ran more seminars and workshops than the local junior college's whole catalog Bob Noyce, Intel50:08.
  • The CUDA textbook sold tens of thousands of copies and was used by hundreds of schools Jensen Huang, Nvidia50:15.
• Find the buyer with a real problem, not the buyer who sees a toy Akio Morita, Sony34:25
  • Sony's first tape recorder flopped when pitched to ordinary people as an expensive toy. Morita became the merchandiser and demonstrated it to the Supreme Court, where stenographers were scarce Akio Morita, Sony34:25.
  • Sony sold 20 machines to the Japanese Supreme Court right away Akio Morita, Sony34:25.
  • Regency released a pocket radio first and quit with little marketing. Sony built the market, even giving salesmen shirts with larger pockets Akio Morita, Sony37:40.
• Fix the image that blocks buyers by selling the category, not your brand Soichiro Honda, Honda36:44
  • Motorcycles were seen as the vehicle of delinquents. Honda advertised in Time, Life and network TV, showed students, housewives and businessmen on Hondas, and aimed at parents too Soichiro Honda, Honda36:44.
  • One mother told a dealer she would not allow a motorcycle, but she would buy her son a Honda Soichiro Honda, Honda36:44.
• Let people try the new thing for free or through someone else's doors Frank Robinson, Coca-Cola25:00
  • With little capital, Robinson and Pemberton mailed free Coke tickets to every address in the Atlanta directory. High-margin syrup made giveaways cheap Frank Robinson, Coca-Cola25:00.
  • Woodruff pledged every soldier a 5 cent Coke. Coke built 64 portable bottling plants abroad and called it the greatest sampling program in history Woodruff, Coca-Cola1:41:11.
  • Schultz spent no marketing dollars. He put beans in Costco, took the United Airlines deal and opened inside Barnes & Noble to reach people with no good coffee Howard Schultz, Starbucks1:09:40.
  • Coke sales went from 600 gallons in 1887 to nearly 10,000 in 1890 Frank Robinson, Coca-Cola25:00.
  • At least 5 billion bottles went to troops, opening markets Coke estimated would have taken 25 years. By 1950 a third of profits were international Woodruff, Coca-Cola1:41:11.

2.5 Remove what stops buyers from saying yes

A new market needs supply, channels and legal room in place, so build the missing pieces or find the gap in the rules.

• Build every missing piece so you can answer yes to a buyer Irwin Jacobs and Andrew Viterbi, Qualcomm1:27:10
  • Carriers could not switch to CDMA unless towers, handsets and chips existed. Qualcomm formed 51/49 joint ventures with Nortel and Sony and made its own silicon Irwin Jacobs and Andrew Viterbi, Qualcomm1:27:10.
  • Fujisawa wrote to 18,000 bicycle shops, promising to help them sell and repair the new motor-driven bicycle Soichiro Honda, Honda29:05.
  • Qualcomm later sold those businesses once the standard had momentum. Revenue grew from $383 million in 1995 to $814 million in 1996 Irwin Jacobs and Andrew Viterbi, Qualcomm1:27:10.
  • Honda had fewer than 200 distributors before the letter and a network of 5,000 dealers after it Soichiro Honda, Honda29:05.
• Check the rules and find the legal gap before betting Irwin Jacobs and Andrew Viterbi, Qualcomm1:14:00
  • Jacobs and Viterbi went to Washington to confirm carriers could use any technology that met the performance spec. Europe mandated its standard, the US did not Irwin Jacobs and Andrew Viterbi, Qualcomm1:14:00.
  • Saul Price saw that a club closed to the public could skirt minimum resale price laws. FedMart sold only to federal employees Jim Sinegal and Saul Price, Costco22:56.
  • Knight hired Steve Prefontaine for $5,000 a year with no duties to get around amateur rules against paid endorsements Phil Knight and Bill Bowerman, Nike1:37:45.
  • The first FedMart did $3 million in sales in 1954 against a hoped-for million Jim Sinegal and Saul Price, Costco22:56.
  • CDMA was officially adopted in 1993, after half the industry had signed Irwin Jacobs and Andrew Viterbi, Qualcomm1:14:00.
• Start where the slate is clean and prove it in the hardest setting Irwin Jacobs and Andrew Viterbi, Qualcomm1:23:45
  • Other countries had no installed analog or TDMA gear. South Korea mandated CDMA for its first digital networks Irwin Jacobs and Andrew Viterbi, Qualcomm1:23:45.
  • When skeptics said San Diego was an easy wireless environment, Qualcomm demoed in Manhattan three months later Irwin Jacobs and Andrew Viterbi, Qualcomm1:22:40.
  • South Korea was close to 40% of Qualcomm revenue for a time. CDMA took 57% of US 2G and 100% of South Korea Irwin Jacobs and Andrew Viterbi, Qualcomm1:23:45.
  • The Manhattan demo won 9x Mobile and later Ameritech Irwin Jacobs and Andrew Viterbi, Qualcomm1:22:40.
• Pick a beachhead, then widen, and wait until you have earned a hard market Jeff Bezos, Andy Jassy and Amazon Web Services, Amazon Web Services1:22:40
  • AWS made startups its first beachhead with free credits and evangelists. It added lift and shift and compliance work for enterprises later Jeff Bezos, Andy Jassy and Amazon Web Services, Amazon Web Services1:22:40.
  • Schultz held off on Italy for decades. He entered only after building the Milan Roastery as a theater of coffee Howard Schultz, Starbucks2:38:20.
  • Netflix moved to AWS in 2009, and Reed Hastings endorsed it at the first re:Invent in 2012 Jeff Bezos, Andy Jassy and Amazon Web Services, Amazon Web Services1:22:40.
  • Straight espresso is now the number one beverage at Starbucks in Italy Howard Schultz, Starbucks2:38:20.

2.6 Where they differ

The right tactic depends on how proven the need is, how long you can survive, and whether the wave is already moving.

• Lead the public, or follow customer signals
  • Morita rejected market research and led buyers with new products Akio Morita, Sony43:45. Jobs also ignored the Homebrew crowd's reaction Steve Jobs, Apple13:54.
  • Kroc did the opposite. He chased a stream of customer calls to the McDonald brothers' stand Ray Kroc, McDonald's0:59.
  • Southwest sat in between. It learned by trial and error, filling a plane that had to fly anyway Herb Kelleher, Southwest Airlines21:05.
• Be early and patient, or be obvious and crowded
  • Nvidia spent a decade in zero billion dollar markets Jensen Huang, Nvidia46:46. Chang survived a couple of years on unstable overflow orders until fabless startups appeared Morris Chang, TSMC1:14:14.
  • Too early can cost you. Microsoft and Google built platforms a decade ahead, and each lost roughly five years of the market while Amazon ran largely unopposed Jeff Bezos, Andy Jassy and Amazon Web Services, Amazon Web Services1:30:00.
  • An obvious bet draws a crowd. Nvidia's 1993 graphics card idea looked like a brain dead yes, and about 90 other startups made the same bet within months Jensen Huang, Nvidia24:30.
• Invent the wave, or ride one already moving
  • Bezos said Amazon would invent or not enter, and refused to copy Apple's music play Jeff Bezos, Amazon38:59.
  • Kamprad rode Sweden's housing boom, which absorbs many mistakes. IKEA was simply in step with an accelerating development Ingvar Kamprad, IKEA42:59.
  • Being first alone is not enough. Regency had the pocket radio market to itself and gave up, and Sony took it Akio Morita, Sony37:40.