Why The Most Dangerous Way to Innovate is the Most Effective Way
Often, the best ideas often look very similar to the worst ideas
In 1963, Seymour Cray’s team at Control Data, based in Chippewa Falls, Wisconsin, built the world’s first supercomputer, the CDC6600.
At the time, IBM was the Google of computers, except more so. Imagine Google, Apple, Microsoft, and Amazon rolled into one. That was IBM.
Its CEO, Thomas Watson Jr., was apoplectic. IBM at the time had hundreds of engineers working on its supercomputer. Control Data was an upstart who built a computer that ran three times faster than IBM’s flagship product. It was akin to Google being displaced as the largest search engine overnight by a tiny startup.
Furious, Watson wrote his famous “Janitor Memo,” in which he said,
“Last week, Control Data … announced the 6600 system. I understand that in the laboratory developing the system there are only 34 people, including the janitor… Contrasting this modest effort with our vast development activities, I fail to understand why we have lost our industry leadership position by letting someone else offer the world’s most powerful computer.”
Hearing of the memo, and Watson’s question about how such a small team could have beaten such a large team, Cray is purportedly said to have replied, “I believe Mr. Watson has answered his own question.”
Small Companies and Innovation
How did this happen? Why is it that across industries, small companies seem to innovate more than large ones? Why is it more likely that Google will eventually be displaced by a small startup rather than by Facebook or Amazon?
Or stepping back a bit: are small companies really more innovative? Is it true or is it a myth?
This post will try to shed some light on those questions. But first, we must understand “The Most Dangerous Equation.”
The Most Dangerous Equation
Let’s say you wanted to figure out what causes kidney cancer. A reasonable question to ask might be, “which counties in the U.S. have the highest rates of kidney cancer?”
The answer is that rural, sparsely populated counties have the highest rates. You might think that perhaps this was due to pesticides, or lack of access to healthcare, or some other factor related to the rural lifestyle.
However, if you were to ask which counties have the lowest rates, you would find that rural, sparsely populated counties also have the lowest rates. In fact, the counties are often adjacent. See below. The red counties have the highest rates of kidney cancer, and the teal counties the lowest.
What’s going on?
Well, when you have only a few people in the county, the likelihood that there will be very high or low rates, due simply to chance, is high. For example, if there were only 2 people in the county and 1 person got cancer, that would be 50%. If 0 out 2 got cancer, it would be 0%.
This is why the best (and the worst) hospitals in the country or the best (and the worst) places to live often are small hospitals and small towns. Statistically, the smaller the sample size, the greater the likelihood of seeing an outlier.
This phenomenon was discovered by de Moivre, and made famous by Wainer’s article, “The Most Dangerous Equation.” Not being aware of this concept can indeed be dangerous. For example, based on the observation that the best performing schools in the country were small schools, a large foundation funded a program to divide big high schools into smaller schools (not large classrooms into small classrooms, but large schools into small schools). The program failed, and only later did they realize that the worst schools in the country were also the small schools.
Size and Innovation
So, turning back to business, let’s take a conundrum that has plagued many industries, including the pharmaceutical industry: difficulty innovating by large companies. Since I know pharma, I will use that as the example.
Big pharma companies have an innovation problem. They only get about 95 cents back for every dollar they invest in R&D for small molecule drugs after accounting for cost of capital (see studies by McKinsey and Deloitte). Yes, the return on investment is below the cost of capital. Below is the NPV graph from the McKinsey report.
It’s puzzling, because big pharma has brilliant scientists, hires the best managers that money can buy, and spend amounts of capital that would drain blood from the faces of managers from other industries. What’s going on? (I should add as an aside, this large company innovation problem is echoed across industries. A friend of mine who works for a leading soft drink company told me that they have decided that they are just unable to invent new drinks and that it was more effective to just buy smaller drink companies.)
How then does big pharma survive? One of the ways is that they buy small companies that have innovative products. There are structural distortions in the industry and human psychology that allows big pharma to systematically buy small biotech at artificially low valuation and (probably) extract more value from the drugs (such as the fact that biotechs with success tend to systematically overestimate their future success rate and therefore over-invest in R&D, which I will discuss in a future post).
Despite this, there is still a controversy in the drug development industry as to whether small companies are actually more innovative. Most people (but not all) believe that innovation comes mostly from small companies, and if you look at where the best selling drugs come from, small companies are over-represented.
On the other hand, if you look at average productivity across the industry, there is pretty decent data showing that the average productivity is the same at large and small pharma/biotech companies. Some people argue that small companies are more productive is an illusion, a narrative fallacy.
Which is it? Are small companies more innovative, or are they not?
As is so often the case with important “either/or” questions, the answer is “both.”
On average, small and big companies are probably equally innovative. But the big companies, because they are so big, are average when it comes to innovation. The small companies are much more likely to be either really good at innovating or terrible at it.
The key factor to realize is that in drug development, and probably in most innovation-driven industries, only the top 5% of product candidates are successful. Being pretty good at innovating is not good enough. It’s a race with only a few winners. The 80th percentile drug is a failed drug. So is the 90th percentile drug.
Drug development is a low-yield sport. It’s like the Olympics. What matters is not the average speed people in a country can run, what matters is how fast the top few runners can run.
And to rub salt into the wound, while 5% of the drug candidates make it to market, the top 1–2% of drug candidates are where almost all the profits come from.
So it’s like this. On one side you have one company with 100,000 employees. The company does good science. On the other side, you have 1,000 companies with 100 employees each. Some of the small companies are doing blindingly good science and some of them are doing cringe-worthy science.
Only the top 1% of the drug candidates win. Which side are you going to bet on? The side with 1,000 average drug candidates or the side with 1,000 drug candidates that range from terrible to brilliant?
Weakest Link vs. Strongest Link, or The Fallacy of the Average.
Some industries and situations call for a weakest link approach. In those businesses, the goal is to avoid mistakes. A company in a lower technology business with a dominant market position is in this type of business. When I worked at P&G, they were very risk-averse. And they were right to be that way. They had a near-monopoly position in many of their businesses, and the only thing they had to do was to not screw up. If they made one mistake and let someone take the pole position, they had a big problem on their hands.
P&G was very, very good at the weakest link problem. Their market research was so extensive and sophisticated, the average marketing manager knew more about statistics than the average medical director in the pharmaceutical industry. The decisions were agonized over to wring every drop of risk out before being implemented. Here is an example of their mindset: the PR guy there once told me, “my job is to keep P&G out of the news at all costs.”
As an aside, there is a debate about whether we, as a country, should be thinking about our society as a weak-link problem. Malcolm Gladwell makes an argument about how we might better put resources to work by focusing on the weak links in his Revisionist History podcast.
How We Got the Dodge Viper
But many businesses are strongest link businesses.
When Bob Lutz, the legendary automotive executive, started his new job at one of the major automobile manufacturers, he was taken aback by how the company selected car designs. The market research group would survey customers and the models with the highest average score would get the green light.
This might sound like a reasonable strategy to most people, but Lutz in his genius immediately knew this was the worst possible way to select car designs.
He said, “Almost no car captures more than a tiny percentage of the market. If you have a car that most of the customers hate but 5% of the customers love, then you have a winner.” He realized that he was looking at a fallacy of the average, and that manufacturing 20 models that were each loved by a few and hated by the most was a superior strategy than making 20 milquetoast models neither loved nor hated by anyone. As Herbert Bayard Swope said, “I can’t give you a sure-fire formula for success, but I can give you a formula for failure: try to please everybody all the time.”
Lutz correctly recognized car design as a strongest link problem.
The Most Dangerous Way to Innovate is the Most Effective Way
Drug development, and most other innovation-driven businesses, are strongest link problems. What drives success in drug development are the 1% best products.
The problem is that often, the best ideas often look very similar to the worst ideas. Both types of ideas sound crazy at first, and it is easy to distinguish between an average idea and a strange idea, but it’s very hard to distinguish between crazy-good and crazy-bad. So, if you want the successes, you have to take more risk. The most dangerous way to innovate is the most effective way. And over the long term, it’s far more dangerous not to take the risk.
This is not easy. The R&D Head at one large company once told me, “this technology looks so promising. I wish I knew for sure it would work, then I would definitely pursue it.” He didn’t pursue the technology in the end. A VP at another company was famous for voting against every single project because he knew that he would be right 95% of the time. If he voted for even 10% of the projects, he would be wrong at least 50% of the time.
But it is possible to take risks at even larger companies. I will discuss some of the ways for larger companies to act like smaller companies in future posts, but let me provide one example.
Genentech is well known for being one of the most innovative biotech companies. It was able to innovate even when it grew to several thousand employees. Its values reflected the risk-taking ethos. Dick Brewer, the former SVP of Marketing once told me, “you can’t build a biotech company not to fail, you have to build it to succeed.” And Sue Hellmann, former President of Product Development at Genentech and now CEO of the Gates Foundation — and one of the most beloved executives in the industry — used to say that if the hair on the back of your neck isn’t standing up when you approve a program, then you’re not innovating.
Unfortunately, risk aversion is very natural when you’re in an industry like drug development where 95% of drug candidates fail — especially if each failure costs $200MM, $500MM, or even more. After couple of dozen drugs fail after burning through hundreds of millions of dollars each, it is natural to think, “we need to reduce risk if we’re going to be successful.”
That’s wrong thinking.
In a strongest link industry, if you have too many failures, you need increase the amount of risk you take, not reduce it. You need to increase the beta — the variability. Yes, you will have more failures, but you will have more successes as well.
To subscribe to email updates, please click here.
“You do not merely want to be considered just the best of the best. You want to be the only one who does what you do.”
— Jerry Garcia
“Be crazier.”
— Masayoshi Son
