On Fast and Slow Failure and Success

I knew only the most basic business strategy of Tech Pro. This was a company that my parens started in 1983 that much later I owned with them and my cousin. The reason I knew so little was that I was only 9 years old. The reason I knew any at all was that I didn’t have to go far that year to learn about Tech Pro. The room, formally a playroom of sorts, had become the “Tech Pro Room”, as it was home to a new dedicated telephone line,  a computer with 64k of RAM and surprisingly a modem to order parts. Our garage was a workshop where my father took in the first of a series of old instruments from shuttered factories, with the purpose of restoring and computerizing them. The kitchen table was not only where I ate my breakfast, but where I worked on my first start-up job, doing Quality Control testing of Tech Pro instruments. The company was financed through seed money from my parents and a small loan from my grandfather. My parents have pointed out that while this seed money was not a lot of money, it was all of the money they had. This money went directly to buy these old instruments. Within the first year of business the company was modestly profitable. Within the first 5 years, the business was comfortably supporting my family, and had managed to computerize a small industry.  Tech Pro was never large, but Tech Pro was international. It employed a great group of talented engineers, many who did not have college degrees but did have creativity and dedication. Tech Pro took me and my family around the world, and introduced us to most of our friends for 25 years. We then sold the company in 2008.

All of this is likely of little interest to you, as Tech Pro is not Hewlett Packer, Apple or Google. It was just a family business that was able to see technological trends early, and create a thriving business based on them. What is interesting to me is to try to understand why this model is not a particularly common one anymore, and where it falls in the spectrum of “expert” advice from tech industry gurus. My friend and fellow board director of our new company (Nanotronics Imaging) Gerry Ohrstrom,  trying to describe our new business and me in particular said  something like “Matthew is a bottom up guy mostly, but also kind of a top down guy too”, which he of course knows. By bottom up he meant that the company was funded initially by us and through revenue like Tech Pro was.  My father is again my partner. We did follow a bit of the Tech Pro example, but this time I felt that we should bring in outside investors (Gerry included), hence being  top down too. It is too early to tell, but this seems to be a nice mix, though not as clear as the one my parents used with Tech Pro. The goals for the new venture are much larger, and therefore the perceived risk and the cost are greater. At least this story provides me with a good justification for being bottom up and top down. I have though been wondering if this is at all common anymore. It used to be. HP, GE, Intel and most early Silicon Valley companies worked in somewhat this fashion. I told the story of BF Goodrich here, where a man with an idea, some patents, a little money and some helpful investors started a 19^th Century business that would become global and very large. The corporate histories of DuPont, Monsanto and countless others of that era are very similar. Despite the overwhelming odds, these companies are still in existence. So why is this model now not considered normal?

I was recently invited to a lunch to discuss an idea a friend has for a science business. I wasn’t the only one invited, but it was clear why I was, since I am the CEO of a science business, not too dissimilar from what the friend wanted to start. The problem with these kinds of meetings that involve any form of advice  is that I am in the middle of the process myself and therefore it is very difficult to have a prospective from the outside, even if it is about someone else’s idea. I have extreme myopia, which I break best from when I am talking about businesses that are completely unrelated to mine. That said, I tried to remember common wisdom about launching a business, which isn’t terribly difficult to do at this moment as I am in the midst of this myself.  I consider my bottom up  and top down approach somehow very old fashioned, and yet very uncommon at the same time. It is not documented as neatly as other models by most contemporary start-up strategists that I have read (though I don’t read all for sure.).  I first assume the things everyone assumes. That is the validity of the disruptive idea (yes, despite saying that I don’t get my advice from gurus, I did read and liked very much the Clayton Christiansen books), the inventions themselves, the basic competitive landscape and all else that seems common sense now.

There is though a view exemplified by the almost cliché concept that since it is not as expensive to start a business as it used to be, due mainly to cloud servers and rapid open source development platforms, everyone should do it. (As a bit of a side note, this is really only true for certain companies. It is still expensive to start companies that make complicated "stuff" rather than those that rely on internet resources only). As the risk is small, the thought is that there is very little downside to trying. Also, because of this a young person can try many ideas in one lifetime, as failure is less risky. Though he may not have been the first to say it, the blogger and prolific book author Seth Godin famously wrote “fail fast and fail cheap. Fail in a way that doesn’t kill you.” This advice was repeated in a recent conversation between Nassim Taleb and Daniel Kahneman, where Taleb stresses this as a fundamental strength of Silicon Valley start-up culture. He was, I believe, thinking of this as a risk mitigation plan. From the Taleb perspective this makes perfect sense. It is true that most businesses fail, and even if the odds were better it would be unlikely that the odds would be so good as to eliminate luck playing a critical role in the success or failure of a business. He is right in this very rational way, and if I wasn’t looking at this from a lifetime on the inside of start-ups, I would take the same rational view. The problem though is not one of expectations from the outside but the one from the inside. It is that position my parents were in in 1983 and that I am in now. It is also the position that more young people ever in technology are in, and from that perspective, every day that we wake up and think about failing fast, or even failing cheap we are failing in a much more profound way than Godin or even Taleb are suggesting. We are failing to risk what we should be risking. For every customer who buys a product that might become obsolete from our failures our failure as a business is a failure to them. This is true for every employee, no matter how much the employee knows she is taking a risk by working for a start-up. We are also failing the investors, even if those investors expect that they may lose the money.

When I think about the cheap part of Godin’s statement, I am reminded by the how little Tech Pro cost to start. Still, it never felt to me like failure was even a mildly uncomfortable option. It seemed dire. Even though my father had excellent career opportunities if Tech Pro did fail, it is hard for me to imagine him thinking about this much as he installed instruments that his customers were relying on. The stakes were too high. Though this is a very personal thing to recall, and may very well have lead me to large enough psychotherapy bills over the years to finance a start-up itself, I remember my mother’s panic the Christmas that Tech Pro was started. They gave me my dream present, an Atari 2600 video console, but not without instilling a feeling of dread which hung over the machine. Mom said “I hope you enjoy this, because Dad just quit his job to start Tech Pro. We may not be able to buy you such a nice gift again.” This comment was not a joke, as her tears were not tears of laughter, but rather it was a statement of commitment. She wasn’t saying this because she felt that Tech Pro would fail. If Tech Pro failed fast and failed cheaply my father would have gotten another job right away, and the following Christmas I would have had another nice gift. Instead the passion became an innovation, the innovation became a company, the company became a responsibility and ultimately that responsibility became success and created far more for far more people than just more Christmas presents for me.

So as I write this I wonder if there is any point of disagreement with contemporary common wisdom. What I come up with is that failure is of course a strong possibility, and that possibility should not stop someone from starting a business. That same failure though, no matter how cheap it appears can never be thought of in those terms. There is a ripple effect in all of the lives we touch when starting a business. That ripple can capsize boats, or guide them like an explorer’s ship to new lands. There is nothing more exciting, but if done right it is both slow and expensive. It is a lifetime that we are lucky enough to likely fail at.

Smart Money?

There is one thing that is true in both start-up technology firms, and in university research labs. In both cases it is hardest to raise money when you need it the most. For the entrepreneur it is also the most expensive time to do it, so many of us wait as long as we can. At least this is what entrepreneurs say to each other, not necessarily wanting to admit that either money isn’t as easy to find as we had hoped, or that we are just too controlling to want to deal with investors looking over our shoulders. To potential investors however, we often say something different. We say that we are not just interested in money for our ideas; we want instead “smart money”. This seems like a perfectly reasonable concept, if not one that is slightly pretentious and condescending. After all, if an inventor needs the money, why not have a smart person give it? You then have positive feedback, at the same time as necessary cash. When this is considered a bit more deeply though, the kind of “smart” a technologist would want in an investor, is not the kind of “smart” he would want in an engineer. For the seasoned serial entrepreneur, confidence may outweigh any interference from having too many opinions on a technological solution, but for most of us it has a way of slowing down a project. For every smart opinion I get (and many are truly brilliant), I stall the progress of my own experimentation and invention, because it only seems natural to want to try good ideas that others suggest. This is almost a requirement if those ideas are coming from the people financing the company.


Does this mean that I want “stupid money” investing in a company I care deeply for? Of course not. If good ideas are a potential hindrance, bad ones are disastrous. What I want are strongly psychologically intelligent investors and partners, who see the big picture of an idea, which requires steps which may be less than perfect, but result in ever improving product releases. For me this is a new type of smart money, and one that is definitely worth pursuing.

Nature's Intellectual Property

I just watched the Charlie Rose interview with Liv Ulmann and Kate Blanchett about the new production of “A Streetcar Named Desire”, playing at BAM in New York. Liv Ullman, a terrific actress herself, is the director of this play, and she said something that strangely applied to the rest of my days conversations. She spoke about that moment when observing an actor that you are directing, when the actor gets it exactly right. The moment is so moving that words cannot describe why it is so perfect. The director is left with a dilemma. Should she tell the actor that he or she got it right, and try to figure out why, or just leave it and be happy it is there? There is risk in both of these approaches. Speaking of it may intellectualize a purely instinctual and brilliant act of the subconscious mind. On the other hand, not speaking of it may mean that it was simply one moment, which may never again be repeated. I know what she is talking about, both from my days in theatre and now working in science. There is a sense of the complexity of inspiration that is humbly rooted in our knowledge of all we don’t know. Human psychology and character development are so deeply intertwined with life experience, theatrical experience, and character interpretation. There is this same thing that happens with invention, and can be equally as fragile.


There are different strongly held beliefs in how to handle intellectual property. Invention of a new technology is not entirely different from the process of bringing a character to life. Like the play, the invention is a unification of previous ideas. Views on how to handle these ideas have varied, and distinguished inventors have disagreed on whether the patent system is truly the best place for them to be revealed. There is an idea, that until the open source movement in software, seemed quant. Benjamin Franklin said after inventing the open stove; “as we enjoy great advantages from the inventions of others, we should be glad of an opportunity to serve others by any invention of ours; and this we should do freely and generously.” This is a highly romantic ideal that has not been very practical. Even not for profit Universities and hospitals now routinely seek patents in order to finance further research. Patents actually do half of what Franklin was suggesting. They do allow others to make and understand the exact invention. They just can’t do it freely for 20 years. By the way, Thomas Jefferson agreed with Franklin on this account. Luckily for them they made their money in other ways, not relying on science and technology for an income. Most private inventors and corporations don’t have this benefit. There is another way that inventions are handled in modern society, which is through trade secrets. The concept of trade secrets is best known in the food and beverage industry. The secret formulas to Coca Cola or to KFC have been famously guarded. This is true though in nearly every product and process, even ones who have extremely strong patents. The truth about trade secrets may be much less brilliant, but more mysterious than a patent. I feel that in most cases a trade secret is something in a process that makes a product unique, even if the company or inventor doesn’t know what it is. I think that it is very possible that Coca Cola does have a secret recipe, but that the recipe by now must have made its way to competitors. The only explanation then on how Coke is still different is that something in the way they make it is different, so they keep making it the same way. This is not so much invention, but chance.

Nature works in similar ways to the trade secret method. There is no patent on trees or minerals. They have come into their present form through a process that worked to keep them intact. Recently while working here in Paris with a very renowned polymer chemist, we were discussing a strange natural phenomenon. For 75 years chemists have been able to create a synthetic rubber which has the exact same chemical structure as natural rubber that comes from the Hevea tree. This was a major development, but strangely when we look at the properties of the natural rubber and the equivalent synthetic, the results are different. With all of our technical and analytical knowledge, we don’t know why this is. For this reason Natural Rubber is still used for many applications. When I was discussing this with my father, he suggested that this was somehow natures “trade secret”. He is right. There is something that for the last billion or so years has been refined to create the latex that is so unique. Nature is not an intelligent being, so likely it does not know why. It just happened, and continues to happen the same way over and over. The same thing is true of silicon, which has a near perfect structure. We would love to create something this perfect in a lab, but we haven’t had the billions of years of trial and error yet. I think with the prototype to evaluate, we should be able to do it faster.

This brings me back to the actors, to instinct and to chance innovation. Perhaps most of what we do is about freezing a process on stage, in a factory or in a lab at the exact right moment. It is also possible that this ability to know when and how to do this is what makes great directors, inventors and companies.

Experimenting

Just because something doesn't do what you planned it to do doesn't mean it's useless.

Thomas A. Edison

If you are an experimental scientist, your days are likely to be either incredibly frustrating, or incredibly exhilarating. Actually for many of us this oscillation of emotions is the natural bipolar state of the work that we are driven to do. Everyone has a slightly unique process for experimentation. I tend to start with improvisation, while other, more organized scientists begin by systematic preparation. An improvisation is by its nature different than an experiment. It is more like psychoanalysis, with free association of ideas, without any conscious direction. I remember this being called brain storming in business and school meetings. For me an improvisation can clear my mind, so that I can see what is already in front of me, rather than be trapped by outside thoughts. As I said though, this is not really an experiment. An experiment requires more than improvisation, it requires an idea, or hypothesis, so that a proper test, and set of testing conditions can be designed. In cases like the Large Hadron Collider at the CERN labs in Switzerland, 15 years have been spent preparing for experiments. One of the key experiments at LHC has been sculpted by the world's leading physicists over much of this time. The Hypothesis is that a unique particle, called the Higgs Boson, can be detected by colliding protons at high energies near the speed of light. Most physicists expect this particle, called by many the "god particle", to be detected, confirming one of the 20^th century's most famous, yet improvable theories in particle physics. This is what is generally thought of as experimentation. At the 2009 Origins Conference in Arizona, two physicists Laurence Krauss and Brian Greene talked of an even more rewarding, or exciting possibility. Dr. Green said "what would be even better than finding the Higgs at the LHC, is not finding it. It would show all of us that there is something else to be discovered. Of course this wouldn't be good for financing another large experiment like this."

Greene was on to something that is generally misunderstood about scientists. Even when an experiment is well planned, and a hypothesis well formulated, we are even more enthralled by the possibility that the experiment leads us to entirely new places. The reason for this is that we trust that nature is inherently more interesting than we can first imagine.

Small technology companies are no less of an experiment than one run in a lab. Like the scientist in the lab, the entrepreneur is putting all of his mental capabilities into a hypothesis, believing that his idea is of value. The good entrepreneur, like the good scientist, is even more moved by the idea which he didn't have. In other words when the experiment of trying an idea fails, he assumes that it must mean that there is an even better solution. This can make for difficult days, quarters and years, but ultimately the openness to reinterpret the experiment can lead to more beautiful places than the original design.

One area of the start-up which is often misrepresented, or at least not thought of in this light, is staffing. When I was a theatre director I was given a common piece of advice which is that "90% of the director's job is casting." This is true of course for directing and hiring engineers, but it is not as rigid as might be implied. When the director Mike Nichols hired Dustin Hoffman for "The Graduate" his choice was mocked throughout Hollywood. Hoffman was too old, to small and too Jewish. The role of Benjamin Bradick should have been given to Robert Redford, by all of the loose metrics of casting wisdom. Nichols was participating in an expensive Hollywood experiment, and one that ultimately paid off with one of the most successful films of its era. In hindsight Nichols is seen as a genius for this decision. When asked about it though, he doesn't see it this way. He claims that the reason for choosing Hoffman was not based on an imagined box office success, but rather just because he thought Hoffman was good. He chose to experiment on Hoffman, not knowing for certain how he would fit in the role, but believing him to be a good enough actor that somehow he would.

"The Graduate" casting example is exactly what plays itself out when hiring the first few engineers in a company, and probably everyone after that. Sometimes it is not always best to hire the MIT Ph.D. with a specialty in your field. Sometimes that is like casting Robert Redford in "The Graduate". It would work probably, but it might not be as inventive as you would like. There was also one other small advantage to the casting of Dustin Hoffman, which at first may seem like a compromise. Hoffman was an unknown, and was not as expensive as Redford. I don't think this was Nichols reason for casting him, but in the end it didn't hurt either. Because the production was under budget in casting, they were able to reallocate some of that money towards the scenery, which included the famous modern and post modern monochrome homes of the Bradicks and the Robinsons. It also didn't hurt Hoffman, as he is now one of Hollywood's top paid actors.

The early years of Tech Pro were much leaner than "The Graduate" pre-production days, but there were some similarities. My parents were looking at doing something that shouldn't have been able to be done with a small amount of investment capital. They were trying to open a software, and hardware company to create completely new technologies, in order to compete with Monsanto, which was at the time a Fortune 50 company. Although it is obvious that this experiment was a risky one, and that there would be challenges, the challenge of hiring seemed easily approached by following common business wisdom: if a company has only a few dollars, at least those dollars should go to the obviously most qualified person. But, what if there aren't even enough dollars to work with, or if it means changing your financing model in order to raise additional funds?

The story of Jeff

The summer of 1985 was a period of transition for Tech Pro. The company itself was experimenting through improvisation and hypothesis, starting as a garage refurbishing shop. That is really all it was. My Father had worked in Akron, the rubber capital of the world (at least then), in many areas of the industry, from manufacturing rubber, to working in a testing lab, to working for Monsanto, who made testing instrumentation. During this time period Monsanto had a near monopoly on a type of instrumentation called rheometers, which were the only practical way of evaluating vulcanization of rubber. In the mid 1980's there was a transformation in the industry occurring. The large tire companies were being acquired by foreign firms, and local factories were being closed. At the same time smaller US companies were filling some of the gap left by the departure of the major players. These smaller companies couldn't afford the rheometers that were by natural supply and demand standards expensive from Monsanto. My parents made a logical bridge between the factory closures, and the need for low cost instrumentation. They purchased used instruments from shuttered plants at auctions, and rebuilt them to resell to the new companies needing cheaper instruments. This was, not surprisingly, welcome news to the industry. It was also a lot of work. Tech Pro hired first a night maintenance man from Kmart to help with the rebuilding. Joe Bulman was a superb tinkerer, and even though hired mainly as a technician, showed creative interests, and abilities. So, he became a design partner, and was the first person to design, along with my father, an original rheometer, not just a refurbished old one. Actually, I will digress for a moment on this story, as it is a perfect example of an experiment that needed adjusting.

In 1985 Tech Pro was actually happy, and even profitable in its business of refurbishing and reselling rheometers. With Joe building, my mother doing the administration, and my father doing sales and installations, it was a nice, very small business. The way the process worked was that Tech Pro would find the old, usually not functional instruments, at an abandoned factory and cheaply acquire them. They would then strip the instruments to only there bare physical structure. They would buy all new parts, from Monsanto, rebuild the instruments, paint them, test them and resell them. This was the entire business at the time. Then a shock that could have stopped Tech Pro at this stage happened. Monsanto refused to sell Tech Pro any more parts. Since Monsanto was the only supplier, there were no other choices. That is except the one that now seems obvious. Tech Pro started to make its own instruments. At this point my father moved from being a salesman, and installation man, to a designer, and Joe became an engineer.

Once Tech Pro had an equivalent, but less expensive instrument to Monsanto, the idea of being just a second supplier lost its excitement. My father was an experimenter at heart, and wanted to experiment with the most exciting technology of the day, the personal computer. Personal computers in 1985 had started to find their way into corporations in many ways. The large main frames of the past were no longer necessary for many applications. Spreadsheets and word processing were being used by nearly everyone. Accounting departments and human resources were starting to use personal computers. In the rubber laboratory, however, analogue devices, called recorders, were the only way to acquire information from rheometers. Personal computers seemed like a perfect fit. A computer would be able to acquire data from the instrument, and store the information. It should also be able to do mathematical calculations to help with the interpretation of that data. The problems in pursuing this line of experimentation were: 1). Joe, my mother and my father had never programmed before, and 2). Computer scientists were scarce and expensive.

Though Joe was the only engineer at Tech Pro in 1985, the work load for building and rebuilding instruments had increased to the point where some hourly employees were necessary to help with the manual labor involved. When a company is as small as Tech Pro, every hire is important, and risky, no matter how unskilled, or low paid the job appears to be. My parents even had a test, which was not so much based on knowledge but instead based on problem solving and creative manipulation. An example from this was putting together a pizza box quickly. Another involved an aspect of design. The only knowledge based questions were ones of electronics. It was important that every early Tech Pro employee know some basics, as everyone needed to multitask. There was also a search for a computer geek. For people who spent time building their own computers, and coding video games. Tech Pro was looking for people who had fun with computers and electronics, not people who were educated in them.

One of these early shop hands was Jeff Archer. Jeff was in his early twenties, high school educated, and clumsy with tools. In such a small firm, where the ability to use a broom, and a drill were more important than your ability to do differential equations, this could have been a problem. Instead months went by with Jeff working hard, but not extremely effectively as an assistant of sorts to Joe. Jeff's potential during this time was growing, as he was indeed leaving work to build his own computers, and doing programming. When my father decided that he wanted to create the first PC applications for rheometers, he did not search for capital, and computer scientists, he instead looked for the geek with the broom. Jeff, and my father worked together to make the first ever PC driven rheometer system. Jeff was not just good technically, he was creative, and smart, and understood, like my father did, the psychology of the user. Together they created a system which was such a smooth transition from analogue to digital, that within 5 years the entire industry had embraced it.

Jeff was an experiment that paid off for him, for Tech Pro, and for the rubber industry in ways that were never hypothesized when he was hired. Still the flexibility and insight to see in him as a potential partner made something unique possible. Only in a small company where the owner knows the worker can this discovery be made.