Ask most successful entrepreneurs how they came up with the idea for their business, and you'd likely learn that what they initially set out to do is very different from the company that you're familiar with. PayPal started out as a service to beam money through Palm Pilots, while YouTube was originally a video dating site. The truth is that early stage ventures are all about experimentation and iteration. As soon as it's written, every business plan is wrong. Good entrepreneurs recognize this, and tend to build agile teams that can quickly respond to early market information in order to identify a real business model and minimize risk.
A necessary side effect of all this experimentation is that most startups will ultimately fail. While the mythical "90% failure rate" has been disproven, I would venture to guess that for technology based startups the failure rate is still extremely high. That's just the nature of the early stage venture world, and ideally it allows the entrepreneurs involved to apply their hard-earned lessons towards more productive ventures. Or, as Jeremy Liew aptly put it: "Companies die, founders and employees learn from the experience and move on, and hopefully start more companies. I for one would love to see the second acts from the teams that are newly freed up."
Today, thanks to a well-documented shift in the online landscape (decreasing storage costs, open-source software, offshore development) it costs much less to start a software-based company than it ever did. Indeed, when I co-founded Infonautics (in 1991) we spent $5M to get our first product to market. At Half.com (1999) we spent $2.5M to get our first product to market. At TurnTide (2003) we spent $750K to get our first product to market. And at Jingle Networks (2005) we spent $300K to get our first product to market. And, in fact, of the 30+ investments First Round Capital has made over the last few years, our average initial investment size is $300K.
Recently we've been seeing a lot of attention paid to startup failures. Techcrunch has even established a “deadpool ” – reminiscent of the old Fucked Company website during the first web boom. This has led some people to speculate that the increased rate of failure is proof that the current funding model is flawed. I disagree.
Although the aggregate number of failures may seem higher (due to the increased number of companies being launched) the ratio of early stage failures to successes is probably still the same. What has changed is that you can now fail faster and cheaper than ever before. While I'd much rather invest in a company that succeeds, if a company is going to ultimately fail I'd rather it fail quickly.
I believe that the goal of seed funding is to validate (or disprove) an entrepreneur’s hypothesis, and thereby “de-risk” the opportunity. Early stage companies should raise enough money to allow them to iterate - as long as their initial hypothesis is still valid and they are making demonstrable progress towards lowering risk. Today’s model of failure is far more capital efficient in allowing entrepreneurs and their investors to do this than the old model. Companies used to waste millions of dollars of VC money – and entrepreneurs used to waste years of their lives – working on a failed hypothesis. Now, the cycle is much shorter.
At Infonautics back in 1991, we raised seed capital to conduct market research – we spent about $150K in surveys, focus groups and secondary research to validate our market. Today, companies can actually launch a product for that amount, in a much shorter period of time. What would you rather see, the results of real market feedback, or the results of market research?
The following graphs (courtesy of Benchmark's Peter Fenton via Venturebeat) illustrate what he describes as the "increasingly Darwinian environment for Internet companies":
Traditional Funding Model
In the traditional funding model a company's risk was decreased by hitting certain milestones, which brought about a corresponding jump in its valuation. Additional funding was needed at each milestone, resulting in a high overall level of financial investment by the investors, and time investment by the entrepreneurs.
New "Cheaper" Funding Model
In the new funding model a startup is able experiment/iterate over an extended period of time for very little capital. Only once some of the venture's risk has been eliminated through accelerating adoption does the company raise more money to further refine the model and expand. Overall the time and cost between the founding of the company and knowing whether both the entrepeneurs and investors should continue to pursue the opportunity is greatly decreased.
I expect that as a seed-stage investor, I will have a much higher number of failed investments than later-stage investors. However, I also expect that I will invest less total capital in failed ventures. Will I be proven right? Only time will tell. For now though, I’d much rather back companies that are able to fail (or succeed!) cheaper.
Thanks to Mazen Araabi for helping with this post...