Cash Flow Forecasting Isn’t What It Used To Be

I share Fred’s determination never to miss out on a later round investment in a company we have nurtured because we do not have adequate reserves. For me, however, the more interesting insight that came out of the exercise of planning our reserves was that this process is different and in some ways more difficult that it used to be.
In the past, most venture backed start-ups in the Information Technology space had a common trajectory. It took a year to eighteen months of development to bring a product to market, another year to establish distribution channels, and then a couple more years to get to sustainable profitability. The capital requirements to fund this trajectory varied depending on the nature of the product, but were reasonably predictable within each segment.
It was also fairly easy to forecast liquidity. Roughly one third of the investments would deliver the vast majority of your returns. One third you’d get your money back. And one third you’d lose your shirt. If the technology never worked you’d stop funding after the initial development phase. If it worked but was undifferentiated and late, you’d stop funding after the second phase. Rarely would you go deep into the third phase in a company that was ultimately unsuccessful.
So across a portfolio of 15 -20 companies, you could model expected cash flows reasonably accurately over the life of the fund. Today it seems harder to do that.
We have talked a lot in this blog about the fundamental transformation taking place in our economy as a result of commoditized information technology and ubiquitous connectivity. We think this is a very exciting time to be investing in early stage web services. We believe that a number of important and valuable companies will be created over the next few years, and we expect to be a part of some of them. But these companies are different. They are more capital efficient. Their success is often built on a network effect rather than proprietary and defensible technology. It is, as a result, harder to predict their capital requirements. It is even hard to predict liquidity, even across a portfolio of 15- 20 companies.
Del.icio.us is an interesting case study. We certainly expected to have more than one opportunity to invest in Del.icio.us. But when Joshua received an attractive offer to sell the company during the Series B fundraising process, he weighed the incremental dilution of a Series B against the likelihood of greater value in the future and made a choice. He was in the position to make that choice because Del.icio.us had proven its value to its users and built a defensible network with very little capital. If it had taken $2-3m in development funding before launch, he would have been facing different economics.
Another factor in this equation is a realistic assessment of the synergies possible through merging a company like De.icio.us with an established portal. In the old days, it was very obvious that if Cisco bought a switch company and dropped it into the bag of each of their sales people around the world, they could dramatically increase sales over night. It is not as obvious that a web service that has already achieved a defensible network effect benefits in the same way by being integrated into a portal.
So what does all this say about cash flow forecasting for an early stage venture capital firm that focuses on web services? It says that it is going to be very hard. The capital efficiency of web services means that few experienced entrepreneurs are going to take a large investment before testing their services in the market. Once the services have proven themselves the entrepreneurs may choose to sell early rather than take the dilution of subsequent rounds. On the other hand, if there are few obvious synergies to be gained by combining a successful web service with an established web portal, they may be very interested in a deal which gives them some liquidity but allows them to remain independent and keep an option on a bigger upside. So in some of our portfolio companies, we may not be able to get as much money to work as we’d like and in others we may be able to invest more than the company needs to fund operations. And there is no way to decide which companies are which up front.
At Union Square Ventures we raised a smaller fund because we believed that it would be to our advantage to have the flexibility to invest smaller amounts initially. We have tried to adapt our fund to the opportunity that exists today for early stage IT oriented venture capital rather than to change our focus. In a comment on Fred’s post on reserves, Bobby asked “ Why wouldn’t you build in a ‘reserve capital call option’ in your LP agreements”. He is touching on an interesting point. We have already adapted our investment strategy and our fund size to the changing market we operate in. Bobby is asking if it will ultimately change the way venture capitalists structure their agreements with their investors? I am not sure we know yet the answer to that question, but we do know that forecasting cash flows is a different exercise than it used to be.