The 1990s Telecom Bubble. What Can We Learn?

My goal when I started writing this blog post was to write a review of The Startup Way. In this new book Eric Ries writes about the application of Lean Startup principles to established businesses and organizations. Rather than write a dry book review, I will try to tell a story in this blog post about a phenomenon I lived through called the Telecom Bubble in order to illustrate a few points that Ries makes in his book. The principal character in this story is a company known as Level 3, which was in the beginning an internal startup inside a huge construction company named Kiewit.

As background, while the Telecom Bubble was related to the Internet Bubble, it was actually a different phenomenon that started in a different way and ended at a different time. How much capital was lost by investors in the Telecom Bubble is impossible to calculate with certainty. Om Malik wrote a book on this period in which he estimated that $750 billion vanished when the telecom bubble burst. That’s a reasonable estimate in my view. The most interesting questions about a phenomenon like this are always: Why did it happen? What can we learn?

Most people believe the starting point of the Telecom Bubble was the purchase of Metropolitan Fiber Systems (MFS). The buyer was a construction company based in Omaha named Kiewit run by Walter Scott. Scott is a very successful business leader who is arguably as famous in Omaha as Warren Buffett. Scott’s ability to build things and otherwise get things done is legendary. With the help of a talented executive named James Crowe, MFS quickly became one of the nation’s biggest providers of fiber based communications operating as a “competitive local-exchange carrier” or CLEC. In 1993, MFS became a public via an IPO. The business model of MFS was relatively sound since it engaged primarily in cream skimming enabled by clever arbitrage of a broken regulatory system. An article in Barons described the situation at the time:

The idea was to offer bandwidth-hungry businesses a money-saving way to route their voice and data traffic past the Bells and other local monopoly providers. The Bells, of course, loathed Crowe and his renegade unit, MFS Communications, for skimming the cream off their customer base. Reminded of this during a recent interview, a chuckling Crowe says, with a refreshing touch of malice: “As if the Bells haven’t been cream-skimming for years by concentrating their efforts on the affluent metropolitan markets, while relegating the remote rural areas of the U.S. to independent phone carriers.”

The mid-1990s was an interesting time in the business world since the Internet was just emerging as a major phenomenon and investing opportunity. As an example of how under the radar the Internet was then, Steve Jobs met with Craig McCaw sometime in the early 1990s and explained to him the nature of the business opportunity that the Internet would create. Craig said to Steve: “Let’s buy it.” Of course doing that was not possible, but people did understand that there were related places to invest that would benefit as the Internet developed. For many people investing in CLECs seemed to be a good place to start. Unfortunately there were many people who thought that same thing at the same time.

Fortune magazine described what happened next:

In April 1996, MFS bought UUNet for $2 billion, a record price for an Internet company. Meanwhile, in Mississippi, WorldCom CEO Bernie Ebbers must have watched mushrooming MFS with some alarm. MFS looked to be a step ahead of his burgeoning empire. Ebbers had a solution: Buy MFS! In late August he offered an astonishing $14 billion of WorldCom stock for MFS. “People made a bunch of money on MFS,” says Omaha investor Ron Carson “people began to think about it like you could get rich overnight.”

The $14 billion dollar acquisition of MFS by WorldCom kickstarted the great CLEC explosion. What was a modest cream skimming business for MFS became a gold rush. Scores of other CLECs were created in just a few years. A partial list of CLECs would include Allegiance, Covad, Northstar, Teligent, Electric Lightwave, ICG Communications, Intermedia Communications, 360networks, Broadwing, Global Crossing and Level 3. The telecommunications business empire of Craig McCaw was no exception creating a company called NextLink (later renamed XO Communications). The problem NextLink and others CLECs faced was overcapacity. Craig McCaw said to me often during this period: “You are always at the mercy of your stupidest competitors.” And in this case there were scores of competitors.

Level 3 was the most glamorous company involved in the Telecom Bubble since Crowe was the leader and because of its affiliation with Scott and Kiewit.  For example, when Level 3  went public the stories read like this one:

NEW YORK — April 1, 1998. In an unusual debut in the Nasdaq Stock Market today, an Internet-related company will begin trading as Nasdaq’s third-largest telecommunications firm, behind giants WorldCom Inc. and MCI Communications Corp. Unlike other recent Internet start-ups such as ISS Group Inc.–which staged a stunningly popular initial public stock offering last week–this newcomer, Level 3 Communications Inc., is far from an IPO. It already has a market capitalization of $9 billion.

How did the vast bulk of telecom bubble end? The Economist magazine wrote in 2002: “George Gilder in February 2001 predicted that two telecoms firms, Global Crossing and 360networks, ‘will battle for worldwide supremacy, but in a trillion-dollar market, there will be no loser.’ Yet within a year, both had filed for bankruptcy—a graphic illustration of how quickly the industry’s titans have been toppled.” The end was shockingly swift. In January of 2001 it was possible for a telecom firm to raise billions of dollars in the bond and debt markets and yet by April it was not possible to raise 5 cents those markets for a telecom business. It took a while for the CLECs to fail depending on how much cash they had on hand. Wen the new cash spigot from the bond and debt markets went dry, it was over. One press report from the time put it simply:

“Once investors noticed that little chance existed for some CLECs to raise the money they would need to continue expanding, an avalanche of selling began. Winstar closed on Monday at 34 cents, off its 52-week high of $54.13, while Teligent finished at 40 cents, down from a peak of $55.50. In essence, both have lost 99 percent of their value.”

A few weeks after George Gilder had said there would be “no loser” in telecom, all of the CLECs were decidedly losers.

You might ask: What about Level 3? The were lucky in that When the financial markets  essentially closed to telecom businesses Level 3 had massive amounts of cash on hand to build networks which Crowe and his team were able to use instead for operations which allowed then to  ride out long painful recovery. In other words, money that was raised to build out new Level 3 capacity was usable for operations during a time when the debt and bond markets were closed. But even then, the road for Level 3 was very, very long and the financial returns modest. Crowe and his team at Level 3 did impressive things to ride out the period and eventually generate an exit with the sale to CenturyLink. People who do what Crowe did with Level 3 don’t get enough credit. Sometimes just surviving something like the Telecom Bubble requires extraordinary skill and effort.  This can be seen in the Level 3 chart that looks like this:

LVLT

The remainder of this blog post will discuss how the CLEC/telecom bubble story, most notably the stories of MFS and Level 3, might be applied in the context of The Startup Way by Eric Ries. Can we learn anything by revisiting the Level 3 story?

All quotes in bold below are from the new book by Ries.

  1. “A value hypothesis tests whether a product or service really delights customers once they begin using it.” (page 93) 

In The Lean Startup Ries wrote: “The value hypothesis tests whether a product or service really delivers value to the customers once they are using it.” This is a critical factor for any business since most products fail for the simple reason that customers don’t value the product enough to pay for it. This may seem obvious, but if it was easy to see whether product/market fit existed there would not be so many failures in the business world. Andy Rachleff, who invented the terms “growth hypothesis” and “value hypothesis,” describes the second term as follows:

“A value hypothesis is an attempt to articulate the key assumption that underlies why a customer is likely to use your product. Identifying a compelling value hypothesis is what I call finding product/market fit. A value hypothesis identifies the features you need to build, the audience that’s likely to care, and the business model required to entice a customer to buy your product. Companies often go through many iterations before they find product/market fit, if they ever do.”

MFS had a proven business model based on “cream skimming” but Level 3 really didn’t have one before it decided to pursue a growth hypothesis. Why? So many CLECs suddenly appeared that the business of cream skimming based on regulatory arbitrage  was just to small to support them all. So everyone running a CLEC at that time grabbed on to a model that was trying to take advantage of the arrival of the Internet. Unfortunately, what looked like a glamorous Internet business was really a business involving backhoes digging up streets, high up front costs and expenses and brutal competition.  Many years later Level 3 would regain a decent business model based on a proven value hypothesis after the other CLECs went though bankruptcy, but it was long painful road. Financial returns today in that business are modest and it scales poorly.

2. “An internal startup is an atomic unit of work in the context of extreme uncertainty.” 

In his first book Ries defined a startup as “a human institution designed to create a new product or service under conditions of extreme uncertainty.”  The simplest way to think about Lean Startup is as a process that applies the scientific method to challenges faced by businesses and other organizations. The Lean Startup is at its core based on a simple idea: since all businesses face inevitable extreme uncertainty the best way to discover genuine innovation is through experimentation. Innovation is discovered rather than planned. What is unique about the business environment today is that experiments cost so little to do as a result of the rise of cloud services and can be completed far more rapidly and with better results since customers are connected to the supplier by electronic networks.

Ries describes the ideal “internal startup” structure in this way:

“A small, cross-functional team with a scarcity mentality, with what we call metered funding, that has a fixed resource budget to apply, that’s accountable to a board of directors. It’s not governed by your usual corporate review process, which is just not at all suited to innovation. Building a minimum viable product using a system we call innovation accounting to measure results, etc. It’s like a special kind of tool in your toolbox, and companies like Amazon do this all the time.”

In The Startup Way Ries argues that large businesses and non-profit organizations can successfully create “an internal startup.” Some of the CLECs including Level 3 were internal startups created inside far larger companies like Kiewit and eventually spun out as independent entities.  Successfully creating an internal startup is tricky since the purpose of a business that has a proven business or operational model is to execute on that model. The unfortunate truth is that operating business and organizations tend to create antibodies against what an internal startup must accomplish to be successful. Senior management must be willing to go the last mile—which can means something that competes with existing business. For the internal startup obtaining the necessary “go to market” resources can be non-trivially hard. Ries outlines a process that he believes will accomplish this goal even inside a business that is executing on another model. Level 3 was able to do this but the teams led by Jim Crowe doing this inside Kiewit had just hit a home run with the sale of MFS to Worldcom. One can argue that this previous MFS home run made the internal startup at Kiewit too easy.

  1. “Long term growth requires a method for finding new breakthroughs repeatedly.” (page 9) In order to creates cycles of continuous innovation and unlock new sources of growth, companies need to have teams that can experiment to find them.”  

Paul Graham of Y Combinator talks about the importance of growth in this way: “A startup is a company designed to grow fast. Being newly founded does not in itself make a company a startup. The only essential thing is growth. Everything else we associate with startups follows from growth.” Existing businesses that have proven business models want to grow too, but that is easier to say than actually do. Many business of all kinds today have dedicated “growth teams” that are responsible for driving new sources of growth. Ries succinctly describes his thesis for achieving growth here:

“The fundamental idea is to treat everything a start-up does as an experiment. Everything a start-up does should be a test — a hypothesis. You really want to organize your company so that it’s built to learn. The Lean Startup idea is based on lean manufacturing—a management philosophy that we can easily adapt to the start-up culture. A key part is creating a feedback loop: build, measure, and learn. We want to get through that loop as quickly as possible. But it’s not just failing fast, it’s failing well.”

The other point that Ries raises concerns the need for continuous innovation. “One and done” innovation simply does not apply in the real world of business. There are too many well-funded and well-run competitors for an business to feel like that can coast along without generating repeated breakthroughs.

In order to grow quickly Level 3 created a state of the art rail plow that allowed fiber to lid alongside train tracks very quickly in relative terms. Unfortunately for Level 3, competitors like Qwest acquired or created rail plows too:

Right now, an army of 1,500 latter-day railroaders is slaving to make the mammoth project a reality: 16,000 miles at an average 120 miles a week. The vast majority of the cable is being laid alongside existing railroad tracks, installed by four or more “rail plow” cars. A custom-designed 25-foot arm extends from those cars and digs up the earth 12 feet from the side of the tracks. It then lays in protective tubing – 2-inch-diameter high-density polyethylene conduit – and cables 4 feet below the surface.

People loved the plans of companies like Level 3, NextLink/XO, Global Crossing and Qwest to build these new fiber networks. Promoters like George Gilder somehow convinced investors that  vastly more supply of fiber back-haul would create unprecedented profitability. Unfortunately for Level 3 (and XO which acquired a 1/3 of Level 3’s long haul fiber capacity) increases in supply kills value without a moat. It is true that vastly greater supply it can create new value in complementary markets like software, but that is easier said that done. Skill in building fiber networks is not translatable into writing innovative software. The amount of fiber and conduits made suddenly available was staggering. When supply explodes, it was not good for prices. Yes, the internet was increasing demand but not as fast as supply was being added in long haul markets. I remember like it was yesterday the night I discovered that Internet capacity was not increasing as fast as UUnet/Worldcom was claiming. This was the key sentence in the paper written by Andrew Odlyzko that I read late one night in August of 2000:

In the intermediate run, there would be neither be a clear “bandwidth glut” nor a “bandwidth scarcity,” but a more balanced situation, with supply and demand growing at comparable rates.

My reaction after reading Odlyzko’s paper that night was: “Holy crap. UUnet is lying.” What I said was similar to the reaction I had when I read about a deal between Enron and Blockbuster for internet bandwidth since I knew that Enron had no bandwidth to sell. Again, “Holy crap.”

The supply limitations that existed at the time of the Telecom Bubble could be found in the metropolitan markets and specifically in the last mile. There were no rail trains to get this work done. This was expensive work that involved getting permits and digging up city streets. These metropolitan fiber build outs ate cash at prodigious rates. The cash burn was ugly and so was the outcome for investors.

4. “A growth hypothesis tests how, given some customers, it’s possible to get more.” (Page 93) 

Ries wrote in his first book: “The growth hypothesis tests how new customers will discover a product or service.” This ties to what Rachleff first described: “A growth hypothesis represents your best thinking about how you can scale the number of customers attracted to your product or service. Growth without value to the customer is likely to lead nowhere–or worse, to a big flameout.” The CLECs were able to grow revenue by brute force methods. Large sales commissions and up front financial incentives for customers can result in the sale of almost anything.  But unless the unit economics of the business are right and the business has a way to fund growth, that will not prevent the inevitable disaster ahead. In other words, selling five dollars for four dollars cannot scale for very long no matter how great the hype around the service. In contrast it is a very wise thing to spend fifty dollars to acquire a customer to pay you ten dollars a month for five years.

Inside the CLECs were many people who had once been in the wireless business and they knew the lifetime value (LTV) model well from that experience. Despite the excitement created by the gold rush atmosphere some of these people could see that customer acquisition cost churn was a big problem for the CLEC business model. They knew that bundling services together increased retention so they acquired related businesses like web hosting firms to add to their service bundle. While adding new services to the bundle did reduce churn, the effort would prove to be too little and too late.

5. “Create an experiment to test [leap of faith] assumptions as quickly and inexpensively as possible (minimum viable product- MVP) (page 86)

Ries talks about a MVP approach here:

“Minimally viable” does not mean operating in a sloppy or undisciplined way, building bad code that’s going to result in a lot of technical debt, or ignoring safety or health concerns. An MVP is not an excuse to throw our beliefs about quality out the window; it’s simply an experiment on the way to excellence.  Instead of taking one big swing with the launch of a new product—devoting months to the design of one technical feature or spending years in stealth mode developing a product without evidence that customers want it—it is an iterative approach to learn who the customer actually is, and what’s honestly required to delight them.

The CLECs were trying hard to find more services to sell as part of their bundle in 1999. As a result, a few CLECs created systems to sell hosted web services. This was the equivalent of software as a service (SaaS) today, but it was provided by firms that were called names like “Application Services Providers.” The software was not optimized for operating off the promises of the customers, but the plan was to see if the dogs would eat the dog food. Many things about this effort did not scale well since few things had been automated. It was a classic MVP. Customers liked the hosted services but the costs were too high given the state of the software. Unfortunately, the clock struck midnight in early 2001 and the sources of new cash disappeared, so the MVP was never transformed into a scalable service that could have helped the CLECs. If you run out of cash nothing else matters. Bankruptcy proceeding are not something that one does happily even if it is Chapter 11 and not Chapter 7.

6. “Experiment to lean what’s working and what’s not (validated learning)”  (page 86)

Ries argues:

Regardless of size, mission, or sector, no organization can survive without the ability to adapt continuously.  I believe that ability has to be a structural part of every organization—that companies need a standardized way to test ideas, run experiments, and follow through on the ones that will bring sustainable growth and long-term impact.

One thing that the CLECs learned via experimentation was that some business was not worth chasing. The customer acquisition costs associated with a commissioned salesperson just killed the unit economics on some small accounts. The only way to stop that from happening was to not give commissions to the sales team on very small accounts. 

The ability of the CLECs to experiment was limited in many cases since they were selling services based on infrastructure owned by the incumbent telephone companies. The incumbents had far better data and control over their systems than the CLECs. Competing with someone using a competitor’s facilities is a really hard thing to do. Some people will argue that it is impossible, especially in today’s world where customer data is so valuable. The joke then was that all it took was a screwdriver for a incumbent telephone company to take down the network of a CLEC. As an analogy, imagine trying to run a bakery using the mixers and ovens of dominant company to compete in the business of selling bread.

The challenges of creating growth are not only numerous but particularly challenging in today’s business world. I don’t know anyone involved in a real business today who does not believe: (1) that the pace of business has increased and (2) that the nature of business has fundamentally changed. What is different? The answer to that question involves so many things that it is hard to know where to begin. Jeff Bezos describes one important change here:

“The balance of power is shifting toward consumers and away from companies…the individual is empowered. The right way to respond to this if you are a company is to put the vast majority of your energy, attention and dollars into building a great product or service and put a smaller amount into shouting about it, marketing it. If I build a great product or service, my customers will tell each other….In the old world, you devoted 30% of your time to building a great service and 70% of your time to shouting about it. In the new world, that inverts.” “Your brand is formed primarily, not by what your company says about itself, but what the company does.”  

Another change is that customers are increasingly connected to businesses which generates valuable telemetry data about behavior and preferences. This data about customers, when combined with advanced analytics and artificial intelligence, enables vastly more value to be delivered to customers. When a business “knows the consumer” as a result of this connected relationship the products and services delivered to that customer are vastly better. If your competitor can do this and you can’t math or better that, you are in deep trouble. The race is on to create connected relationships with customers. Everything that can be connected via a network will be connected. A radio or wired connection is going into almost everything. Key to all of this is trust- if that does not exist the data provided by the customers will not be as valuable. Customer trust is earned is an asset that is as important as anything on a balance sheet.

7. “Take the learning from each experiment and start the loop over again (build-measure-lean feedback loop)” (page 86)

What artificial intelligence enables a business to do is conduct thousands of growth experiments which can potentially enable a better products and services. Most of the growth experiments will fail, but that is a natural part of the scientific method that Ries advocates in his books. The vastly lower costs that cloud services enable means  that most experiments fail or even nearly all experiments fail is OK since the payoff from the experiments that succeed can be so massive.

Ries suggests:

  • Enter the Build phase as quickly as possible with an MVP that will allow us to test a clear hypothesis we have about our product or strategy.
  • Measure its impact in the marketplace using actionable metrics that help us analyze customer behavior.
  • Learn whether our original assumptions about the product, process, and customer needs were correct or whether we need to change strategies to better meet our vision.

The Build-Measure-Learn feedback loop does not end once we’ve put our first product into the market. Every new launch of an MVP is an opportunity to gain valuable information about how well we’re meeting customer needs—and whether we need to adjust our strategy with a pivot. 

The CLECs knew that serving building with fiber was expensive so they bought radio spectrum in the millimeter wave bands (e.g., 24 and 38 GHz) to try and connect business using an approach they called LMDS at lower cost. Rather than rolling out the service slowly after product/market fit was proven, the CLECs starting building out these LMDS wireless networks quickly. There was no iterative process where the value hypothesis was proven first based on a MVP. The result was a disaster. Why the rush? Some of it was fear of missing out (FOMO) but some of the rush was driven by the fact that people knew inside the CLECs that the fiber-based approached really wasn’t working well financially and that an answer was needed quickly. It is hard not to fall prey to confirmation bias and loss aversion when the business really needs a solution and has a lot of time, money and reputation invested already.

8. “Because of the short term pressure, [businesses tend to focus] on the projects they believe will maximize that quarter or fiscal year.” (Page 14)

Ries is talking in that sentence about the danger of a short-term focus in business. This short-term focus is often created by dysfunctional metrics that I call Siren Call Ratios (e.g., IRR, P/E and RONA). I wrote about Siren Call Ratios recently in my blog post about the demise of post-break up AT&T. In that example, while the long distance business seemed like be a wonderful opportunity to AT&T based on the Siren Call Ratios, it had zero barriers to entry. The long distance business without a local loop monopoly was one of the most deadly mirages in business history.  If you manage a business to make ratios like P/E look good in the short-term you can end up with a long term disaster.

9. “It is important to create metrics that measure success… that often replace a company’s traditional metrics- often ROI…” (page 176) 

A business that is managed to achieve vanity metrics has never been more at risk. In today’s economy ever business must think and act based on long-term financial returns in order to survive. If a business is instead focused on short-term metrics, competitors that do invest for the long-term will come along and crush it. Customers are far too well informed given the information transparency created by the internet to accept inferior products. One of the better examples of a company that is driven by long-term metrics is a company that has the following operational model for its internal startups: Once a year (in summer at the start of their fiscal year) the company “stack ranks” proposals that product managers believe will drive growth. Every product manager (PM) must make a case for what they want done at this stage. The PMs essentially sign up for improvements to key metrics like revenue and cash flow growth in their six page proposal. The senior leaders of the business then do the stack ranking.  Some things make the cut and some don’t. There is another valuation period six months later. I am making it simpler than it really is since they must write an imaginary press release first, but those are the key steps. Compensation and promotions are based on hitting the numbers in the metrics. Post-mortems are done on failed efforts. Do you know or can you guess who this company I just described is?

10. “Provide metered funding to startup teams… denominated in a fixed budget of either time or money.” (page 289)

More startups, including internal startups, die from indigestion than from starvation. Too much funding can very often be a curse for a  startup rather than a blessing. When a startup is over funded, cash is too often used to solve problems that are better tackled through innovation and organizational culture. Too much funding can also cause the organization to grow headcount too quickly and to work on the growth hypothesis to begin before the value hypothesis is solved.  All of this happened in the case of CLECs.

FOMO on the CLEC opportunity caused legions of business people and investors to bet massive amounts of capital on an industry that would eventually be shown not to have a sound business model. It is worth pointing out that the company that kicked the telecom Bubble off with the purchase of MFS was Worldcom. The New York Times article on that company that is most relevant read in part on July 22, 2002:

“WorldCom, plagued by the rapid erosion of its profits and an accounting scandal that created billions in illusory earnings, last night submitted the largest bankruptcy filing in United States history…. the chief financial officer, had devised a strategy that improperly accounted for $3.85 billion of expenses.”

Over a period of years it became more and more clear that the unit economics of the CLEC business simply didn’t work financially. Here is a short description of the CLEC investing thesis:

The CLECs built a fiber ring in a city and connected that system to some office towers. The landlord must agree to allow the fiber provider and access to a room for equipment, Fiber must be run up a “riser” to each floor. All that requires cash outlay up front. The CLEC then hired sales people who go to buildings and travel up and down the elevators trying to sell a bundle of local phone service, long distance and Internet connectivity. Sometimes they also sold web hosting in that bundle. The lifetime value (LTV) calculation for a small office tenant might look like this:

clecs

That’s an ugly financial outcome that continues linearly with scale. When the end to the telecom bubble came (i.e., when the cash ran out), the demise of the CELCs was swift and merciless.

11. “On a regular schedule (cadence) make a decision about whether to make a change in strategy (pivot) or stay the course.” (page 86)

One of the hardest things about internal startups is that there are no market forces that sort out which efforts should be shut down and which should be funded further. The process can become arbitrary and political in a large organization. By creating a regular review cadence with a set of relatively independent reviewers these intra-company startups/ experiments all potentially face a process that mimics the market. Instead of seeking financing in rounds from venture capitalists internal startups try to convince a board that they have achieved certain success metrics. This is not an easy process to make happen in a large company, but it is a necessary one. During the Telecom Bubble the markets did not perform their usual function due to the power of FOMO (fear of missing out) until billions of dollars had been lost. Obviously the markets were neither perfectly efficient of rational and agents were not perfect informed when it came to CELCs. Level 3 survived until it was recently bought by Century Link, but it was a long painful road.

12. “Modern companies need the ability to consistently and reliably make bets on high risk, high reward projects without having to bet the whole company.” (page 45) 

Amazon is a leading example of what Ries is writing about in that sentence. Bezos describes the objective at Amazon as follows:

 One area where I think we are especially distinctive is failure. I believe we are the best place in the world to fail (we have plenty of practice!), and failure and invention are inseparable twins. To invent you have to experiment, and if you know in advance that it’s going to work, it’s not an experiment. Most large organizations embrace the idea of invention, but are not willing to suffer the string of failed experiments necessary to get there. Outsized returns often come from betting against conventional wisdom, and conventional wisdom is usually right. Given a ten percent chance of a 100 times payoff, you should take that bet every time. But you’re still going to be wrong nine times out of ten. We all know that if you swing for the fences, you’re going to strike out a lot, but you’re also going to hit some home runs. The difference between baseball and business, however, is that baseball has a truncated outcome distribution. When you swing, no matter how well you connect with the ball, the most runs you can get is four. In business, every once in a while, when you step up to the plate, you can score 1,000 runs. This long-tailed distribution of returns is why it’s important to be bold. Big winners pay for so many experiments.

If you find yourself wanting to refresh your memory about Learn Startup by reading my previous blog post about Ries, the unfortunate news is that it is now only available in my new book. About 35 of my previous blog posts have been rewritten in a revised format that has been professionally edited. That book will be available in a few days and can be pre-ordered now. 100% my royalties from that new book will go to the charity No Kid Hungry.

Notes: 

http://money.cnn.com/2001/04/10/technology/herring_clec/

https://www.amazon.com/Startup-Way-Companies-Entrepreneurial-Management/dp/0147523303

https://www.amazon.com/Lean-Startup-Entrepreneurs-Continuous-Innovation/dp/0307887898/ref=pd_lpo_sbs_14_img_0?_encoding=UTF8&psc=1&refRID=V9B2W5YKZ23VS2J01P84

http://www.leanblog.org/2017/10/podcast-290-ericries-leanstartup-startup-way/

https://www.inc.com/lee-clifford-julie-schlosser/lean-startup-eric-ries-testing-your-product.html

http://www.economist.com/node/1234886

http://archive.fortune.com/magazines/fortune/fortune_archive/2002/07/22/326291/index.htm

http://www.nytimes.com/2002/07/22/us/worldcom-s-collapse-the-overview-worldcom-files-for-bankruptcy-largest-us-case.html

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=236108