Tag Archives: pattern recognition

When Investing Too Early Isn’t “Wrong”

28 Jun

I recently had a twitter exchange with early stage venture investor Mark Suster over the issue of being “too early” as a venture capitalist and what that means in the current climate. While I genuinely like and admire Mark and am not one to publicize my disagreements (or would that be disa-tweetments?) with other investors over what are primarily philosophical matters, I thought the content in the exchange was valuable and supported some re-examination.

Mark made what I would consider a somewhat sweeping statement–namely, that investing too early was the same as being wrong. I responded that this entirely depended on the perspective being considered. Let me explain.

There scarcely exists a technology sector that became vibrant and consequential that did not experience a great deal of stops, starts and stalls early in its evolution. Most of the early “failures” in a space — i.e., companies that had a piece of the solution figured out but perhaps not the entire solution — were venture-backed companies. Many of the same venture investors in those early failures learned from the experience and went on to back later entrants that became juggernauts in that given sector. Were those investors wrong? Would those investors have developed the insights, market knowledge and ecosystems critical to their becoming investors in the eventual market leaders without those early experiences? Would the sector have developed as it did without the flame-outs?

My answer to those rhetorical questions is emphatically “no.” As has been evident  across technology markets for decades, technology advancement occurs in waves of innovation that beget other waves of innovation, and so on. The early failures provide formative experiences for the investors that were involved and for future entrepreneurs behind new entrants in that sector that can plainly see what worked and what didn’t for the previous class of market entrants.

Take, for example, the notion of Pattern Recognition among venture investors. Much has been written about the concept, including an earlier piece in this forum. Tangential to pattern recognition is the idea that a venture investor’s early forays into a sector provide a great education — wanted or unwanted — about what will ultimately be successful in the space and what simply doesn’t work. It also provides the investor a great deal of market knowledge and an ecosystem of supporting companies and entrepreneurs that will serve that investor well for years to come in his or her chosen area of investment, and even in other areas. This is called active cross-pollination and it is critical for long-term success as a venture investor.

To pull on that thread a bit further, venture investing is not a discrete, narrow vertical exercise. Good investors are constantly influenced by outside factors, parallel markets, advancements in technologies that appear unrelated until an epiphany occurs and interesting combinations can result. From the embers of a previous failed investment can come a dormant technology or a seasoned manager that can be re-potted into a new venture and can enable that venture to become enormously successful. The point I make in raising this is that the future success would not have likely occurred without the early mis-step. I have several examples of this in my own career.

There is an oft-told piece of black humor among attorneys about a veteran attorney counseling a young protegé. The senior attorney remarks, “When I was a young attorney like you I lost a lot of cases that I should have won. Now, with my years of experience, I win a lot of cases that I should probably lose.” No one is suggesting that venture investors new to a sector should message to entrepreneurs that they are “learning” on their deals. That said, to maintain that experiences on early investments do not positively inform decisions made on later ones is simply folly.

So, can one be “wrong” by investing too early in a sector, as Mark suggests? Most definitely. This occurs when an investor develops an investment thesis, makes an investment in a company against that thesis, leaves the sector after that company fails to never invest again in the space and never leverage the lessons learned from that experience into other investments. Hopefully, this does not occur very often among professional venture investors. The mere statement that anyone invested “too early” in a sector implies that the given sector did ultimately develop into something substantial. With any luck, the earlier investors that helped shape the sector with early bets were able to prosper by participating in the eventual winners. Historical venture returns seem to bear that out.


In Defense of VC Pattern Recognition

13 May

The notion of venture capital best practices is nothing new. For some time now, I’ve enjoyed the back and forth between venture practitioners and entrepreneurs about what constitutes the elements of good process and cogent thinking in venture decision-making. Codifying a venture process or methodology has stymied many a well-meaning firm over the years. While there are certain processes and approaches that most firms would recognize as well-reasoned, there still exists wide disagreement among venture investors and the firms they helm about what should drive investment decisions, how much rigor and analysis should play into investment decision-making versus old-fashioned gut instinct and momentum-style approaches, and what ideal blend of background, experience and temperament makes for a successful venture capitalist. The answers aren’t so obvious.

For me, one of the exciting elements of venture also happens to be one of the areas of greatest frustration — to both those in the field and those outside the glass. That frustrating element is the intangible, unstructured and fickle nature of the venture business. The cliché about venture career paths — that to ask ten VCs how best to enter the venture business would elicit ten very different responses — has persisted for the simple reason that it is an accurate depiction of how circumstantial the venture business is. Hollywood agents may have their William Morris Agency mailroom farm league system to help groom aspiring agents-to-be (and weed out the dreamers), but no such finishing school exists for the Sand Hill Road or Route 128 set. Founding and operating a start-up is probably the most obvious and germane way to learn the mechanics of start-up operations and to develop the skills that, later on, will prove invaluable as an advisor to other startups. That said, a succesful entrepreneurial career is no guarantee of subsequent success as a venture investor. Indeed, some of the industry’s most successful venture capitalists never spent a day in a start-up, so go figure. 

Correspondingly, just as there exists no feeder school to a venture career, there is no tried-and-true approach to hone good decision-making by both aspiring and practicing venture investors. Into this debate the notion of pattern recognition has been both praised and derided. For some it is a valuable tool — some might say skill —  to help investors separate wheat from chaff. It can frame for an investor likely outcomes from past experiences with similar companies, similar approaches, similar strategies and similar team compositions. Detractors would argue that pattern recognition is not really a skill at all, just an accumulation of trade practices and a prism steeped in groupthink through which venture investors view opportunities freed from truly independent thought and creativity. The problem with pattern recognition as a technique to evaluate an opportunity, the argument would go, is that it is not only a bad substitute for independent thinking, it prematurely nips in the bud potentially promising opportunities by unfairly punishing them for the past failures of companies that came before with similar concepts.

To be fair, that’s not a bad point. Sure, the fact that the last ten companies that pitched me trying to do a GPS-enabled location-based gaming app focused on dog groomers came to nothing doesn’t necessarily mean your company won’t be the next Google. Some start-up team, it could be argued, will finally succeed in that space and that team could just as easily be yours — but I wouldn’t bet on it.

From my perspective, both camps are right to a certain extent. Like any tool, pattern recognition can be too heavily relied upon and can improperly take the place of real thinking and analytical rigor. So, to the extent we are discussing deal evaluation at the earliest stage, I would wager that many investors do prematurely jettison opportunities because their perspective is skewed by knowledge of prior stumbles by earlier market entrants. The more nuanced approach, I would wager, is to more heavily weigh the attractiveness of the space the company is focused upon over the company itself in the initial stages of evaluation. If the investor believes wholeheartedly in the space and is not dissuaded by the fact that so many earlier entrants failed trying to build a business there, then the discussion revolves around whether that investor wants to invest in that space and whether that particular company can succeed. Any forthcoming investment decision, or so goes the hope, will be made based upon the merits of the company and its solution and not upon the number of carcasses on the battlefield from prior attempts to “take that hill.”

To my mind, perhaps the greatest value of pattern recognition is simply a function of robust deal flow. Seeing a LOT of opportunities in a given space is invaluable and absolutely critical to long-term success in a particular area. Sure, there always exists the one-off moonshot opportunity that can land in a venture investor’s lap that becomes the next juggernaut in the space, but the more typical scenario behind most venture success stories is one of a methodical investor meeting every company in the space, learning the idiosyncracies of the ecosystem, and then partnering with the company that that investor believes has the right combination of team, solution, and timing to become a market leader. The pattern recognition that develops in that process, obviously, is a function of seeing so many opportunities in the space, weighing their pros and cons, and getting past the initial frisson and excitement of what every company is doing — and, hence, being dispassionate about the investment selection process itself. I find pattern recognition, as a tool, far less effective as a company matures and the relationship with that company deepens. I have been on great boards and dysfunctional ones; I’ve advised talented CEOs that commanded the respect of everyone in their organizations and I’ve worked with CEOs who couldn’t run a lemonade stand. While I have learned from each of these experiences I find these lessons rarely apply cleanly to subsequent problems. Companies, like people, are unique and applying a rote process or management algorithm for specific eventualities based upon how earlier crises were managed rarely works effectively.

%d bloggers like this: