At a somewhat recent dinner with Brad Feld (@bfeld) and some other great Boulder tech entrepreneurs we got into a discussion about Ray Kurzweil’s book on the Singularity. A heated debate ensued whether his theory about the convergence of technological computing capacity, and our ability to understand (and thus integrate computing into) the brain was believable or not. Brad graciously bought everyone around the table the book, and we agreed to meet back in a month and debate it further. Since then I have been plowing my way through the book. Kurzweil is, if nothing else, the Philip Glass of technology theories. He pounds you over the head with example after example of data that supports his arguments. Without going into my opinion on the overall subject (I have about 100 more pages to read before I think I can form a conclusion), it did make me think a bit about the challenges facing the venture capital industry these days. Seth (@sether) wrote a good overview piece of the continued struggle of the VC industry to right size itself in the last few years and find its equilibrium point. I wanted to chime in a bit on this subject from the perspective of what I see as the coming venture singularity.
The starting premise of the singularity is that computation grows
exponentially over time for the same cost (duh!). That means in 10 years
we’ll have 100x the capacity to do what we do now. It’s basically a
Moore’s law analysis across computing, hardware, bandwidth, densities, etc. I don’t think anyone can disagree from the data that computation is on this trajectory, and we’re now hitting the interesting part of the curve (e.g. the knee). What we’ll be able to do in 2020 will be almost unfathomable given how we look at the world today and few of us (me included) will even be able to conceptualize what 2030, let alone 2040 will look like. This is interesting as it will happen in my own lifetime (bummer to those born in 1900).
What underlies the challenge of the VC industry is many companies (let’s stick to pure software for the time being) are seeing an exponential reduction in the cost to build, test, market and deploy solutions. The inertia of the VC world was founded upon the need to spend $2M to build a product, $1M to take it to market and $5M to grow the business to a substantive point ($10M run rate was the traditional crossover point for practical profitability in software startups for the last 20 years). Fortunately for entrepreneurs that has all changed, and thus unfortunately for a number of Venture Capitalists who’s models (e.g. fund sizes) are stuck in the past.
Trada was built on a very different model than previous companies I have started. We intentionally raised a small amount of capital and built without any capital assets (we have only one server to run Exchange which is a vestigial limb of the fact that I couldn’t get over wanting Outlook – I know, I know – sorry guys). The process of development was and still is constant iteration with a set of customers on a live product. We deployed our first version in Dec of 2008, only 2 months after we literally painted the walls in our office. And we have been running live and upgrading live ever since. In the space of only 7 years (from the beginning of the last company I started), we are working with a completely different financing model based on the technology model in which we could build, test, market and perfect a product.
As the cost to develop and deploy has gone down (e.g. incremental costs
through Amazon EC2 et al), the cost of marketing your product has shifted too. Social media provides an unbelievable amplifying affect when you are ready to take your message to the market. Online blogs and media have aggregated millions of potential customers as loyal readers. We saw this with our launch. A mention in a few pubs drove incredible lead generation for us, and we parlayed that into ongoing leads through retweets and blog traffic.
With SaaS/Web2.0 applications there is basically no customer install and no upgrade cost. SaaS companies sell most of their software over the phone (thus no cost for very expensive direct sales folks which take 6 months to pay back the investment you have to make in them – no offense to direct sales folks – its just the truth of those dollar dynamics). And outbound marketing of your product has become exponentially easier. Testing various forms of marketing whether it be paid search, display advertising and social networks can be done incrementally and without major $$ commits. You can quickly figure out what works and invest more heavily into profitable sales through that channel. Once you’ve maxed out one channel, you can find the next and invest in that one. There are about 50 ways to do outbound online marketing now if you count all the ad networks, email, social media sites, etc.
According to Kurzweil’s theory (more broadly applied), in 10 years this whole process which now takes about a year will shrink by an order of magnitude. Every aspect of “launch” from software development to computing to ecosystem engagement (e.g. social networks, twitter, etc..) and traditional lead generation marketing will shrink as well in cost and time. And thus the overall cost to build, launch and grow a software business will shrink as well. If it took 12 months and $1M before it will take 1 month and $100,000. Some companies building iPhone or Facebook apps are already seeing this model in action. The few fringe companies that have figured this out will become the norm in 10 years. So where does this leave a Venture investor?
If an investor can’t deploy capital, they can’t make returns. Some firms will have to shift their model to investing in later-stage companies that have already rounded first and need $5-$10 to go big and try and hit a home run. Other firms will need to restructure their portfolio approach to investing lots of small chunks of money in many companies and getting a lot more smaller wins quickly to achieve the IRR expected in the venture investment class. And for earlier stage firms, they will inevitably have to shift their engagement model as well. Young companies are much more susceptible to making wrong turns and going down cul-de-sacs. Learning is expensive in time and money. Venture firms can fill an incredible hole in this equation though by becoming hyper-operationally focused. My relationship with our one institutional investor, Foundry Group, started based on this premise. They have been phenomenal operational partners and have added incredible value to helping steer a fast moving company through its learning phases more efficiently than if they’d given us twice as much money ,and we’d just spent money to learn.
So all of this is to say that I think Venture firms focused on software are nearing their own sort of singularity: a place where success can be tried and tested for $100s of thousands of dollars and not millions. Trying to deploy $200M of capital across 20-40 companies in this model simply won’t work. The trick then is to ask how the human element of a venture firm can scale. 5 partners is still 5 partners and they have limited capacity to dig deeply into many companies. The most interesting change to watch will be firms that try and figure that part of the equation out.
As a side note, this model is already emerging in a roundabout way
through the likes of TechStars, Y Combinator and other mentor focused organizations. The owners of those organizations take a % of the company, but for the first time institutionally the percent they take is valued partially on the cash and partially on the operational value of the mentors they have. They spread this money and mentorship across 10 companies at a time and play a bigger portfolio approach. Recently TechStars success with this model was highlighted, and it gives us the first glimpse of what right now seems like a nifty idea and in the future may end up being the actual model everyone uses.
Either way, investments always evolve with the market. The venture singularity is coming, and I’m vastly interested to see who takes advantage of it first.
(Cross-posted @ Trada Blog)
