Markets recovered faster than the economy did, which is a familiar pattern but a useful one to write down. This letter is about what we got right in the crisis, what we got wrong, and what we are now doing differently. We have learned, twelve years into this work, that letters written in the immediate aftermath of a crisis tend to overstate the lessons. We will try, this December, to keep the lessons modest in proportion to our actual confidence in them.
On Three Convictions That Were Tested and Held
The conviction that a family-fund structure provides material advantage in a downturn was tested explicitly and held. Several of our peer firms with strong vintages from 2005-2006 produced 2009 results that materially trailed ours; the differential is almost entirely explained by reserve availability. We had reserves; they did not. The structural difference is the kind that compounds across decades, and we have rarely been more confident in the architecture of the firm.
The conviction that founders with personal capital at risk behave differently from those without it was tested across our entire portfolio and held without exception. The four companies we recapitalized last year all had founders who had taken personal positions in the recapitalizations; the two we declined did not. We do not believe this is coincidental. A founder who is unwilling to commit personally during a crisis is a founder who has already, in a sense, decided. The crisis simply makes the decision visible.
The conviction that the right time to commit to a long hold is at the bottom of a cycle, not the top, was tested by the four first checks we wrote in early 2009 and the early performance of those four companies. We will not characterize the early performance in detail. We will say that the four are tracking ahead of the operating plans they presented when we underwrote them, and that the founders' early decisions — particularly around hiring and customer concentration — have validated the diligence we performed.
On Two That Were Tested and Did Not
The conviction that companies with strong unit economics are protected, in a downturn, from significant valuation compression — was wrong. Unit economics do not protect against the absence of marginal capital, and many of our portfolio's strongest businesses traded at marks below what their fundamentals would have justified, simply because the buyers had left the market. The mark-to-market exercise we conducted in March produced lower numbers than we would have predicted; the operating performance underlying those numbers was, in retrospect, materially better than the marks suggested.
This is a useful adjustment to our framework. A company's value is the function of its operating performance and the available capital pool simultaneously; underweighting either factor produces incorrect estimates. We have, since March, been more attentive to capital-pool conditions when projecting future valuations. The shift sounds obvious in retrospect. It was not obvious in advance.
The conviction that we could distinguish, in early 2008, between durable consumer behaviors and cyclical ones — was wrong. We adjusted some positions on this basis and were wrong in two of the four cases. We are humbler about this distinction in 2009 than we were a year ago. The general lesson is that consumer behavior is more resilient to recessions than economic theory predicts, and that adjusting portfolios on the assumption of consumer-behavior breaks tends to produce errors that compound. We will be slower to make such adjustments in the next cycle.
On the Bitcoin White Paper
A nine-page paper circulated on a cryptography mailing list in October 2008, and the network it described launched in January of this year. We did not understand it on first reading. We did not understand it on third reading either. What we understand now, after watching the network operate for eleven months, is that the experiment is more serious than we initially thought, and that the questions it raises about the institutional structure of money are questions our framework was not built to evaluate.
We do not know what to do about this. We are reading. The paper makes a technical claim — that distributed cryptographic consensus can produce a coordination outcome previously thought to require a central intermediary — that is, in our reading, novel and rigorous. The economic claim — that the protocol can support a stable monetary system at scale — is more contested, both in the literature we have reviewed and in our internal discussions. We are not yet positioning around it. We are paying attention in a way we were not in October.
The lesson we draw from the year's reading on this topic is broader than the specific protocol. It is that the most consequential technological developments increasingly arrive in forms our financial framework does not have categories for. We will need to invent the categories. The work of doing so is, we suspect, going to occupy more of our research time over the next decade than any other single topic.
On the Mobile Buildout, Now Visible
The App Store, eighteen months in, has demonstrated something we underwrote in 2007 and did not yet have evidence for — that an always-connected pocket-sized computer changes consumer behavior in ways that are large and lasting. The companies we backed during the crisis on this thesis are, almost without exception, ahead of where we expected them to be. The combination of the iPhone shipping, the App Store opening, and the data-network capacity of major carriers maturing has produced consumer-software environments that no one — including the founders themselves — modeled adequately in 2007.
We expect mobile-native consumer products to dominate venture returns through 2015. We are now positioning specifically around the second-order implications: mobile commerce, mobile advertising, mobile content, mobile communications. Each of these will produce companies that are not direct iPhone-store applications but that depend on the mobile substrate for their operating assumptions. The shape of the next decade's portfolio is becoming visible.
A Closing Note
The most useful work we have done this year has been the public review of our own framework. We will keep doing it. There is nothing more dangerous, in our work, than a conviction that has not been forced to defend itself.
The Partners
Winzheng Family Investment Fund · December 2009