The horse is too small, the jockey too big, the trainer too old, and I’m too dumb to know the difference.
– Charles Howard, owner of Seabiscuit addressing comments about his horse, Seabiscuit and the team of people he had assembled to train and race Seabiscuit. Red Pollard was Seabiscuit’s jockey and Tom Smith was its trainer. ((The Wikipedia entry on Seabiscuit is here: http://en.wikipedia.org/wiki/Seabiscuit))
The debate about whether to “bet on the jockey” or to “bet on the horse” is as old as horse racing itself. Historians can’t pin-point an exact date when horse-racing as an organized sporting spectacle began, but available records suggest organized horse-racing dates at least as far back as Ancient Greece, and Ancient Rome, and probably earlier than that; in China, Persia, Arabia and other areas of the Middle-East, and Northern Africa. In North America, the first horse racing trophy dates back to 1665. ((Encyclopedia Britannica. Accessed on August 30, 2013 at http://www.britannica.com/EBchecked/topic/272329/horse-racing#toc3310))
Venture capitalists have had an ongoing debate on that question too. For the purpose of this post, I will narrow things down to early stage venture investing in startups at the seed, series A and series B stage of development.
For the purpose of facilitating how I think about this later in this post, I will define a startup in a very specific way. A startup is a temporary organization in search of a scalable, repeatable, profitable business model. ((Steve Blank and Bob Dorf, The Startup Owner’s Manual Vol. 1: The Step-by-Step Guide for Building a Great Company, California, K&S Ranch Press, 2012, page xvii)) Once this search is complete, the organization ceases to be a startup. It then becomes a company. How do we know when a startup is undergoing the transformation into a company? It starts to resemble a mature corporation, with the functions, management hierarchy, and organizational reporting structures that we recognize as indicative of mature corporate establishments. I am assuming that the transition from startup to company happens between series B and series C, but there’s a lot of room for error in that assumption.
It is important to observe a few things. Thanks to the work of many people, it is now accepted ((Steve Blank, Why The Lean Startup Changes Everything, Harvard Business Review, May 2013.)) that early stage startups engage in activities centered on search and discovery. Hopefully these activities lead to the discovery of a business model that works. During this period of search and discovery, business plans as we have traditionally thought of them are not very relevant because many elements of the value creating, value delivering, and value extracting activities of an early stage startup change so rapidly that the traditional business plan becomes obsolete before the ink used to print it has dried. ((I am not suggesting that the thinking that goes into developing a business plan is completely wasted. For early stage startups authoring a full-blown business plan is not the best use of resources. That changes once the search for a scalable, repeatable and profitable business model is complete.))
In conducting my research for this post I found an approach to thinking about the horse-racing analogy that I find useful. In this approach there are three elements one has to think about if one decides to bet on a horse-race; the jockey, the horse, and the course. ((Jeffrey L. Minch, A Jockey, A Horse, A Course. Accessed on August 30, 2013 at http://themusingsofthebigredcar.com/a-jockey-a-horse-a-course/)) In this formulation the course is the market, the industry and the economy within which the startup operates. The horse is the business model. The jockey is the founder, or co-founding team, of the startup.
Does it matter in what market or industry the startup expects to do business? Yes. It does.
Profitability is not similar between industries. In a Harvard Business School case, Pankaj Ghemawat and Jan Rivkin compare the between-industry profitability of 15 selected industries starting in 1988 and ending in 1995. ((Pankaj Ghemawat and Jan Rivkin, Creating Competitive Advantage, HBS Case 9-798-062, 1998. Revised December 20, 1999)) According to this analysis pharmaceuticals and software were the two most profitable industries. Motor vehicles and scheduled air transport were the two least profitable. There’s no reason to believe that the overall conclusions of that study would be materially different today. The unique nature of the competitive forces at play in each industry lead to potentially vast differences in industry level profitability. The average investor in a software or pharmaceutical startup could expect better returns than the average investor in a scheduled air transport or motor vehicle startup simply as a result of industry level profitability. They also studied intra-industry profitability over the same period. Here they found that intra-industry differences in profitability could be larger than between-industry profitability. While they found differences across the pharmaceutical companies that they studied, all the companies in that industry created value – they earned positive profits. On the contrary, the most profitable scheduled air transport company earned a relatively large, positive, profit while the least profitable companies in that industry earned negative profits – they destroyed value. Fully 43% of the companies in the scheduled air transport cohort were value destroying. ((The cohort had 7 members, 3 had negative profits.))
Do the kind of business model options that are available to the startup matter? Yes. They do.
In an MIT working paper published in May 2008, the authors studied 10,970 publicly traded US companies in the COMPUSTAT database from 1998 through 2002 and developed a typology of 16 business models. They reached two broad conclusions. First, business models matter. Second, some business models seem superior to others, although none perform better than all others across every performance measure. ((Thomas W. Malone et al, Do Some Business Models Perform Better Than Others? (May 18, 2006). MIT Working Paper 4615-06. Accessed on August 31, 2013 at http://seeit.mit.edu./Publications/BusinessModelsPerformance12July2006.pdf))
Holding business models ((This is not entirely possible, but let us assume so.)) and industry constant, does management matter? Yes, all else held equal management matters. A lot.
Ghemawat and Rivkin found differences in profitability between companies within the same industry cohort which often exceeded the between-industry differences in profitability. For example, in the pharmaceutical industry cohort Johnson and Johnson and Schering-Plough were the most profitable. ((There were 15 companies in the pharmaceuticals cohort.)) They both were more than 6 times as profitable as the least profitable member of that cohort, and at least 3 times as profitable as the second-to-last most profitable member of the cohort. As I have already described the gulf between the most profitable and least profitable in the scheduled air transport cohort is even more startling. In that case the most profitable company in the cohort is about twice as profitable as the least profitable company, on an absolute basis. However, on a relative basis the most profitable company created value while the least profitable company destroyed value by earning negative profits.
The most profitable companies in any industry accomplish profitability by implementing strategies to negate the most unattractive features of their industry, while simultaneously exploiting the most attractive features of the industry. They do that by pursuing unique and value creating activities that cause them to become indispensable to the industry network. Also, to become a market leader a company must successfully manage the tension between the industry structure that it inherits and its position within that industry. Strategic choices are entirely within the purview of management. Management decides what activities to pursue and what choices to make. To a large extent, these choices determine the company’s position within its industry. The most profitable companies in an industry have management teams that develop and implement strategies that lead those companies to be more profitable then their peers.
The two studies I have relied on so far for this discussion did not focus on venture-backed companies. I found one that did. A 2004 paper by Steven N. Kaplan et al ((Kaplan, Steven N., Sensoy, Berk A. and Strömberg, Per Johan, Should Investors Bet on the Jockey or the Horse? Evidence from the Evolution of Firms from Early Business Plans to Public Companies (August 2007). CRSP Working Paper No. 603. Available at SSRN: http://ssrn.com/abstract=657721 or http://dx.doi.org/10.2139/ssrn.657721)) attempts to specifically “address an ongoing debate among venture capitalists (VCs) concerning the relative importance of a young company’s business idea and management team to the company’s success.” The authors studied 50 venture-backed firms from “early business plan to initial public offering (IPO) to public company (three years after IPO). They concluded “that on the margin, VCs should spend more time on due diligence of the business rather than management.”
They also found that the majority of VC-backed and non-VC-backed IPOs do not involve a line of business change at some point during the firm’s evolution from startup to company. For this analysis they studied a sample of 106 IPOs that occurred in 2004. ((They eliminated 200 IPOs because the nature of the companies involved did not allow an easy comparison with the typical venture-backed startup.)) Overall a line of business change occurred only 7.5% of the time. ((There were 106 firms in all, with 8 of them undergoing a line of business change at a median date of 7 years before the IPO.)) In the VC-backed sub-set of that sample, a line of business change occurred 8% of the time. ((7 out of 88.)) In the non-VC-backed subset a business line change occurred only 5.6% of the time. ((1 out of 18.)) When they applied this line of analysis to the sample of 50 startups that were the focus of their main study they found that a line of business change occured only 2% of the time. The data suggests that the vast majority of startups that succeed remain in the same line of business that they were pursuing at the time that venture capitalists made an initial investment. Apple, Microsoft, Cisco, Ebay, Amazon, Google, Facebook and LinkedIn did not change their core line of business between early startup and IPO, though they have each broadened the scope of their activities as they develop new organizational capabilities, and applications for their products.
At this point it is worth recapping what we know from the academic research; First, a company’s profit potential is directly related to the profitability of the industry in which it operates. The race-course at which one chooses to bet on a horse race is important. Second, a company’s profit potential is also directly tied to the type of business model it chooses to implement. Given the race-course, the horse one chooses to bet on at the race is important. Third, a company’s management chooses its strategic positioning, and implements the strategic choices that the company pursues. The best management teams lead their companies to earn a disproportionate share of the profits available to all companies in that industry. Given the course, and the horse, the jockey one bets on is important.
Given the working definition of a startup with which I began this post, it is also worth considering what we don’t know. Ghemawat and Rivkin studied publicly traded companies. I assume that their conclusions apply to early stage startups too, but I cannot be 100% certain. It is more likely that the conclusions they reach in their study become applicable once a startup exits the search and discovery phase of its life-cycle and begins to grow and scale its business. Malone et al studied publicly traded companies. We assume their conclusions hold true for early stage startups but we cannot be certain, since early stage startups experience rapid change that leads to changes in important elements of the business model, or sometimes an entire change in the value-proposition of the startup. That is to say, often a startup sets out to solve one problem. Then its founders realize they will never solve that problem before their startup fails. At that point they opt to change course by finding a new problem to solve. I assume that when such a pivot occurs, the startup will remain in the same industry – a software startup does not pivot into the scheduled air transport industry, for example. Kaplan, Sensoy and Stromberg studied VC-backed startups. But we do not know how early these VC-backed startups were. I do not know if they were still in the search and discovery phase, or if they had just transitioned into the grow and scale phase – in other words, were these seed, series A and series B investments or were they later stage VC investments?
Because of what I do not know about the academic research I have studied so far, I do not feel I should draw any hard and fast conclusions, although that research is helpful in guiding me towards a framework as to how to think about this question.
Some further digging turned up an interview ((The video is online at http://www.youtube.com/watch?v=9sYsKW0LSwA&feature=youtu.be)) between Ron Conway and Paul Graham ((Eliot Durbin, a partner at BoldStart Ventures in New York City pointed this out to me when I told him I was working on a blog post to explore this topic.)) Paul runs Y Combinator and Ron Conway is an angel investor who invests through SV Angel. Ron was an early investor in Ask Jeeves, Google, Paypal, Facebook, Twitter, Twilio, and Square. ((He has invested in more than 650 startups.)) Ron Conway said two things that I find particularly relevant. ((Given that this interview took place in November 2012, it is also worth noting that Ron Conway thinks that we’re in the early days of investment opportunities for investors in internet startups.))
First, the most important decision he has made as an early stage investor was in 1994 when he decided that he would only invest in startups developing internet software. In other words he picked his race-course. Second, he invests in the entrepreneur first. Once he’s decided to invest in an entrepreneur his approach is to invest in every startup that the entrepreneur might create. So he picks a jockey, and makes a lifetime commitment to that jockey.
As Jeffrey Minch suggests in his blog post, this is a problem about permutations and combinations. ((Suppose you have n objects. A permutation of these n objects taken r at a time is an arrangement of the objects in which order matters. A combination of the n objects taken r at a time is an arrangement in which order does not matter.)) Given that early stage investors need to bet on the course, the horse or the jockey, it appears to me that early stage investors will be most successful if they are correct in picking at least 2 out of the 3 most of the time they invest. ((There is only one way to pick all 3 objects correctly – a combination of 3 objects taken 3 at a time, and I am assuming that investors rarely know enough to decide if they can bet on being correct about all 3 characteristics of a startup.)) The trick to solving this problem in a way that maximizes the early stage investors chances of making profitable investments is to reduce this from a choice where one is choosing 2 objects out of 3 to one in which one is choosing 1 object out of 2. That is what Ron Conway accomplishes by focusing on internet software startups only. He has reduced the problem to one in which he has to only bet on the horse or the jockey, because he’s already made his bet on the race-course. He has chosen to always bet on jockeys, given his choice about the race-course at which he is willing to place bets.
So what have I learned? I have learned that if I am going to succeed as an early stage investor I will improve my chances by picking a course and studying that course well. Next I have to take a bet on the jockey by looking for the traits that I think will lead that jockey to succeed. Finally, since every early stage startup is searching for a business model, I have to be willing to work with the jockeys I bet on to identify the right horse for the race. In other words, I have to work with the startup founders to search for and discover a repeatable, scalable, and profitable business model given the problem they are trying to solve.
There is still much that I do not know in relation to this question, but I think this gives me a useful framework.