Summary: Great by Choice describes the results of a deep investigation into how young companies can survive and thrive in chaotic, turbulent environments to achieve spectacular results. The book is of great value startups and entrepreneurs seeking to build enduringly great companies. In this blog post, I look at how his concepts of fanatical discipline, productive paranoia, and empirical creativity apply to building a startup that succeeds over the long-term .
I just finished reading Jim Collins’ new book Great by Choice: Uncertainty Chaos and Luck—Why Some Thrive Despite Them All (GBC from here on out). GBC is the spiritual sequel to a highly-regarded & best-selling book published by Collins in 2001 called Good to Great: Why Some Companies Make the Leap… and Others Don’t. Both are great reads, but I find GBC particularly relevant to technology entrepreneurs (like myself). Why? Two reasons.
The level of research behind the book:
Unlike many business books, this is not just one successful guy waxing philosophical about how he made stuff happen . Jim Collins and his coauthor Morten Hansen had entire teams of research analysts work for 9 (!!) years to complete the book.
They picked industries that were highly volatile and selected young/small companies that did extraordinarily well (beating their industry’s average stock growth by 10x or more for at least 15 years). They found comparison companies that were started off very similar to the “10x companies” but only had average performance, and dissected all the data they could gather on both companies to find the differences. For more, see Appendix A below.
The companies / industries studied:
- Computing/Software: Microsoft vs Apple 
- Integrated Circuits: Intel vs AMD
- Biotechnology: Amgen vs Genentech
- Medical Devices: Biomet vs Kirschner
- Surgical Devices: Stryker vs USSC
- Insurance: Progressive vs Safeco
- Airlines: Southwest vs PSA
The companies are relevant and familiar to tech entrepreneurs like me and many of the folks on this blog. My focus in this post is to look at how the conclusions from the research could be applied to early stage startups that WANT to build enduring and spectacularly successful companies. I’m excited to see what we find.
Myth-Busting: It’s not about more vision, creativity, risk-taking or luck
One of the great things about this study is that it’s not just studying winners but looking at the difference between winners and losers. GBC found that the 10x companies were NOT more creative, visionary, ambitious, lucky, hard working, risk-taking, innovative, etc. It’s not that those things weren’t important – I think they were/are. And GBC acknowledges this.
It’s just that both groups had lots of these things. Yet they had different outcomes. So we have to look at what DIFFERED between the 10x and comparison companies. Let’s start by looking at how innovation happens at 10xers.
Fire Bullets then Cannonballs: Another look at launching MVPs
In the startup world, most people are familiar with the concept of the MVP: the minimum viable product. The idea is that before you prep your product for scale, spend lots of money on marketing/advertising or build advanced features — you first need to find that MVP that really delights customers. GBC confirms that within 10x companies were not necessarily characterized by gigantic innovations/breakthroughs, but an intelligent system of “empirical creativity” that they call “Fire bullets, then cannonballs”.
Basically the 10x companies tended to fire more bullets (testing products/services/channels,etc in a limited capacity), did NOT fire more cannonballs (big deals/major efforts/expensive acquisitions) vs comparison companies in total, but when the 10xers did go all out, their efforts were more likely to result in positive outcomes.
Microsoft In 1987 Microsoft had built Windows to run on the IBM PC based on MS-DOS, which was becoming the industry standard. But IBM was also preparing to release a new set of computers running OS/2 – which Gates himself thought would dominate the industry within 2 years. Microsoft had only IPOed a year ago and still had limited resources. Gates hedged his bets and continued working on Windows development even as he committed tons of resources on developing on top of OS/2 — in the face of strong resistance from some of the people in his own executive team. This turned out to be a really good idea when OS/2 failed to become the market winner. Microsoft was then able to throw the full weight behind Windows and launched Windows 3.1 selling a million copies in 4 months (3x more than OS/2 did in 3 years).
Apple Today, Apple makes boatloads of cash on their 300+ retail stores – earning more per sq ft than Tiffany’s. But it didn’t start that way. Jobs recruited the President of GAP to the Apple Board way back in 1999 and spent a year testing the concept before they launched their first two stores in 2001. Through refining their stores over a decade, Apple now has more people coming through their stores in one quarter than went through all four Disney resorts in 2010. That’s how you go from bullet to cannonball. 
What I think this means for startups:
Don’t feel like you have to be THE first company out there, or the first one with a specific feature. Gowalla was founded either one or two years before foursquare (depending on who you ask) but foursquare won. Yes, it’s valuable to be first for many reasons, but it’s not worth being first on something that sucks.
Look at Zynga: criticize them all you want about Farmville being a ripoff and a waste of time. The company has IPOed, which puts them among a very select group when it comes to internet/consumer tech startups. In the games business, you need hits. Zygna uses “ghetto testing”to make sure that they invest their resources in projects that are going to pay off, based on user data. This is what it means to fire bullets then cannonballs as a startup.
On a personal note, when we launched Ridejoy on the West Coast in the fall, we had already tested a lot of elements of our rideshare service from our month long Burning Man rideshare site BurningManRides.com in the summer. And now we’re taking the lessons we’ve learned from the West Coast to our SF <> Tahoe route.
The 20 Mile March: Great companies are built through fanatic discipline
This concept is about fanatic discipline when it comes to your company’s performance. Specifically, setting high but reachable goals (that are specific to the business and largely within your control) and consistently hitting them, even during bad times, without overreaching or getting greedy during good times. By doing this you build your company’s confidence to perform an adverse conditions, reduce the likelihood of catastrophe when hit with turbulent disruption and helps you exert self control in an out of control environment.
I chose to lead with the empirical creativity section because we’re talking about startups, but the book lead with the 20 Mile March because it was “a distinguishing factor, to an overwhelming degree, between the 10X companies and the comparison companies in our research”. Wow. Discipline is what matters.
The race to the South Pole Jim Collins lives in Colorado and is an lifelong rock climber, so it’s not surprising that he takes inspiration from famous expeditions. One re-occuring example is of the race to the South Pole in 1911: two teams battled to be the first – one lead by Roald Amundsen and another led by Robert Falcon Scott.
Amundsen turned out to be the 10xer of the comparison for many reasons, but for this section we’ll focus on his relentless discipline. Amundsen pushed his team to go 15-20 miles every day despite weather conditions. On good days he would not overextend (even against the wishes of his team) and on bad days he made sure they kept moving forward. In comparison, Scott’s team fluctuated wildly with their march, sometimes traveling a lot and sometimes traveling not at all (with similarly varying weather).
Amundsen hit the South Pole in mid December and arrived back at camp exactly on the day he had scheduled: January 17, 1912. Scott arrived to the South Pole in January, cursed the Amundsen flag, and headed home, ultimately getting stranded in a storm and freezing to death in March, only 10 miles from a supply depot.
Southwest Airlines Southwest had a 20 Mile March of being profitable each year – a demanding task when the US airline industry as a whole turned a profit on only 6 of 14 years (1990-2003). Despite major changes in the industry where big players went bankrupt, Southwest was able to stay profitable for 30 years straight, including in 2001 amidst the 9/11 attacks. It also had the discipline to not grow too quickly even when lots of cities demanded their service, expanding for instance, to 4 new cities in 1996 even as 100 demanded SW. They left growth on the table in order to be more steady.
Stryker (biomed devices) CEO John Brown set a benchmark of 20% net income growth every year in 1977, calling it “the law”. Excuses were not acceptable. Executives behind the 20% mark got the “Snorkel Award” because you were basically going to drown. At regional sales meetings, people who make the 20% sat at the front and the rest sat in the back. The company hit the goal 19 out of 21 years – and also held back, growing more slowly than the comparison company USSC more than half the time.
What I think this means for startups
I think there are two elements to the 20 Mile March: 1) figuring out the fundamentals of your business and 2) sticking with them and making steady, consistent progress.
Most startups are still figuring out the first part, which is why I focused on the empirical creativity section first. You’ve got to figure out a winning formula, but once you do, drive at it hard and consistently, avoiding the overreach but also doing everything in your power to keep forward momentum constant.
It’s difficult to get a current day example of this given how secretive companies are about the way they run, but I’ll point to two examples of companies being relentlessly and consistently focused on acheiving specific metrics:
Facebook: realized that growth was going to the be their path to success and built a growth team designed specifically to work on projects and features (like tweaking the signup flow) to push for growth.
Amazon: in its 1997 letter to shareholders Bezos laid out his plan: win marketshare.  Amazon has consistently grown it’s revenues and its share of markets, (like the book and ebook publishing markets) to staggeringly heights. They are always trying to drive down prices and implement programs that make people buy more, even when it doesn’t make sense from a revenue/cost perspective (like Amazon Prime). Some analysts have indicated that Amazon could be running up to 1/3 of all ecommerce and growing.
At Ridejoy, we’re in it for the long haul. I’m very excited about our business but I know it’s going to take hard work, applied consistently over time to make it work. It took Airbnb 1,000 days before they became an “overnight success” and we’re also prepared for a long march.
Leading Above the Death Line – Why productive paranoia is critical for surviving and thriving
It turns out that 10x companies are not a cocky bunch. Yes, they work toward their 20 Mile March goals with steadfast resolve and they’re confident, but they’re also hypervigilant and aware of the countless number of things that could weaken or destroy everything they’ve built. Leading above the death line is not about completely avoiding any errors, but about making sure you have what it takes to recover and nothing wipes you out.
Specifically, the research found that:
- Cash buffers: 10X companies held a 3-10X higher than average cash to assets and cash to liabilities ratio, and 80% had higher ratios vs the comparison companies
- Made less risky decisions: 10x companies made fewer small, medium and high risk decisions compared to comparison companies
- Deliberately slow or fast: 10x companies often didn’t make the first move and demonstrated a higher level of internal data gathering/analysis. But when they were sure it was the right move, they usually moved faster than comparison companies.
Microsoft Early on, Bill Gates established a rule that Microsoft should be able to go a full year without any revenue. Even in 2010, they are close: having around $40B in expenses and $36B in cash/cash equivalents. In 1991, MSFT stock dropped 11% when a “Nightmare Memo” written by Gates internally was leaked, listing a series of challenges and threats in competitors, technology, IP and customer support failures that could wallop the business. John Sculley of Apple in comparison, had a good year in 1998, doubling sales and net income in 2 years, so he went on a 9-week sabbatical. Not particularly paranoid.
Intel Andy Grove (former CEO of Intel) is the Godfather of Paranoia, having penned an management book called Only the Paranoid Survive discussing how to respond to “Strategic Inflection Points”. Intel of course famously dropped their memory chip business when it became exceedingly clear, after rigorous analysis, that they would not be able to compete with the Japanese and jumped directly full force into microprocessors.
What I think this means for startups:
This chapter immediately reminded me of Chris Dixon’s article “Always have 18 months of cash in the bank“. In re-reading Dixon’s article, I was struck by this line: “building/marketing/selling technology always takes longer than you think”. Basically, shit happens and you need to build a solid buffer for yourself (if you are running a VC-backed startup) to make sure you don’t have your back against the wall (ie your company’s death line).
Additionally, the point about the 10xers taking less big risks reminds me of Steve Blank’s point that startups focus on reducing risk wherever possible, whether that be invention or market risk.
Basically – always be thinking about what could kill you, do your best to prepare / build buffers for it and don’t feel obligated to take a risk unless you’re sure it’s the right decision. And always have a plan b.
Return on Luck: You won’t be lucky, so be good instead
Luck, the dirty word in all success/business/self-improvement literature. Perhaps some people and some companies just get lucky and our ability to draw useful lessons and conclusions from their success is just not possible. Collins and his team anticipated this and devoted an entire chapter to luck.
GBC applied a consistent methodology to both pairs of companies to analyze how luck played a role in their outcomes. About 230 luck events were categorized and studied, each meeting all three criteria of being unpredictable, independent of the actions of key players, and having significant good or bad implications for the business.
Examples of luck events include: Amgen isolating the gene for EPO, which it likened to “finding a sugar cube in a lake a mile wide/long/deep” or the New England Journal of Medicine publishing a paper that challenged the effectiveness of one of Genentech’s major drug products.
What they found: neither 10xers or comparison companies had substantially more good luck or bad luck events, nor did one giant piece of good luck carry a 10x company through all its success. Luck exists but it tends to even out the playing field. What matters is “Return on Luck” or how you take advantage of good luck and avoid choking.
More specific example:
AMD – In the mid 1990’s, AMD had a huge amount of good luck in their direction: computer makers were kind of sick of Intel’s dominance in chip making and wanted an alternative. A federal jury allowed AMD to make Intel chip clones, and thus earned record sales. Then IBM announced they were pulling a bunch of computers because of a flaw in the Pentium chips, forcing Intel through a $475 million recall process.
This is some insane great luck for AMD. People were preordering AMD’s K5 chip like crazy. And what happened?
They blew it. The project slipped months behind schedule and people started moving back to Intel. By the time AMD fixed their problems, Intel had released a new generation of chips, forcing AMD behind again. They weren’t disciplined about their hardware production and failed to build up the buffer to “throw money at the problem” and thus failed to take advantage of the good luck.
What I think this means for startups:
It’s true that many smart people believe luck plays a big role early on, but once you’ve hit some kind of scale, this fact matters a lot less. Yes, luck will play a role in your startup, but luck is not a strategy. Nor is it really something you should spend any time thinking about.
You can’t control luck, so just prepare for the worst by strengthening your team with exercise and strong relationships, and your business with cash, users and a consistently improving product. Then prepare to seize opportunities when they arise and never let go.
Steve Blank calls a startup “a temporary organization searching for a repeatable and scalable business model”. Some of the concepts here don’t translate when you’re running an early stage startup. But many of the ambitious entrepreneurs I know want to build their companies into spectacular, enduring businesses and I think it’s good to look down the road and learn what works.
I have a tendency, as I think many entrepreneurs do, of operating on gut instinct, playing things really up near the wire, and switching my focus when I get bored. Like a mosquito trying to score a sting as Paul Graham so eloquently put it in his essay on making wealth. 
However, as I consider my goals for Ridejoy – to be as successful as Airbnb/Dropbox/Foursquare and perhaps even Google/Apple/Amazon, I realize that what works when you’re two guys with a prototype does not when you’re hundreds of people with millions of users.
Crazy one-off bold moves, strokes of genius or disproportionately good luck aren’t the secret. Instead, it’s discipline, preparation and intelligent risk-taking. Less sexy perhaps, but backed by empirical data.
I’ll close with the final paragraph of Great by Choice, where Collins describes the moment where we are afraid, exhausted or tempted and have to make a choice:
We are not imprisoned by our circumstances. We are not imprisoned by the luck we get or the inherent unfairness of life. We are not imprisoned by crushing setbacks, self-inflicted mistakes or our past success. We are not imprisoned by the times in which we live, by the number of hours in a day or even the number of hours we’re granted in our very short lives. In the the end we can control only a tiny sliver of what happens to us. But even so, we are free to choose, free to become great.
Many thanks to Vicki Mach and Bilal Mahmood for reading earlier versions of this article.
 I think that if you’re trying to found-n-flip a business, most of these lessons do not apply. They’re specifically for founders/leaders who want to be a long lasting business success.
Additionally, I don’t want readers to come away with the idea that these are the *only* ways to become an enduring success. However, we have more evidence to suggest that these ways will work compared to many other approaches.
 Those books can be good if they are written by someone with a lot of humility and a great memory, but usually they gloss over the hard stuff / failures and focus too much on the author’s personal traits.
 It’s funny because everyone thinks of Apple as a success story, not as a mediocre comparison company. And it’s true, from 1997 to 2011, Apple stock grew nearly 6,700%. But GBC started researching in 2002 and so Apple’s stock was pretty crappy given these were the “lost years” when Jobs was off building NEXT and Pixar. However, Collins acknowledges that in the past decade, Apple is much more a model of a 10xer, and indeed has demonstrated the effectiveness of many of these practices.
 “More people now visit Apple’s 326 stores in a single quarter than the 60 million who visited Walt Disney Co.’s four biggest theme parks last year, according to data from Apple and the Themed Entertainment Association.” – Wall Street Journal, Secrets from Apple’s Genius Bar
 I tried really hard to find more specifics about Facebook’s growth team be could not – please let me know if you find anything (when it was created, how many people are in it, what they’ve done, etc)
 The quote specifically: “We believe that a fundamental measure of our success will be the shareholder value we create over the longterm. This value will be a direct result of our ability to extend and solidify our current market leadership position.The stronger our market leadership, the more powerful our economic model.”
 “A startup is like a mosquito. A bear can absorb a hit and a crab is armored against one, but a mosquito is designed for one thing: to score. No energy is wasted on defense. The defense of mosquitos, as a species, is that there are a lot of them, but this is little consolation to the individual mosquito.”
Appendix A: Methodology
Collins’s takes business research to the next level. Here’s a breakdown of exactly how Great by Choice was written. Their research question: “Why do some companies thrive in uncertainty, even chaos, and others do not?”. How they answered the question:
- Picked industries that had a high number of random but powerful and dangerous events outside of the control of the companies in the industry.
- Employed a matched-pair comparison in order to contrast what the winners did and the losers didn’t.
- Studied companies that were small or young when they started their success, specifically ones that IPOed between 1971 and 1990.
- Narrowed to companies that beat their industries by at least 10X over a 15 year or more period
- Selected comparison companies that were in the same industry, and similar age/size to the 10X company, but with only average stock performance.
- Poured through press clippings, historical documents, interviews with staff, annual reports, 10k’s etc
- Both authors separately read all documents, each then wrote 70+ pg case reports on each company + matched comparison
- Based on data + analysis, they extracted practices and behaviors that explain what the “10x companies” did that the comparison companies did not
This rigorous nearly decade long research study makes the conclusions of the book far more grounded in reality. While we cannot claim causality, the chances that the following behaviors are random/unrelated to success appear very low, especially when you consider that almost all the winners did them and almost all the losers did not.