Sixteen years in, the market is hotter than we have seen since the start of our firm. Late-stage mobile valuations are expanding at a rate that recalls 1999 — but unlike 1999, the products underneath them are real. In Beijing, a news application that orders its feed entirely by algorithm has just passed its first year — a kind of product that, two years ago, would not have been seriously discussed. In this environment, we want to use this letter to say what we have learned, what we no longer believe, and what we have come to believe more deeply.
On the Vintage That Is Repricing Risk Off the Table
Late-stage rounds are now closing at multiples of next-year revenue that, two years ago, would have required a company to be either profitable or growing at three-digit annual rates. We are seeing rounds at these multiples for businesses that are neither. We will not participate at these prices. We are aware that this discipline will cost us paper performance over the next eighteen months. We are also aware that the framework which justifies the discipline is the same framework that justified our positioning in 2008.
The structural difference between this market and 1999 is that the underlying products work. The mobile internet has produced a category of consumer behavior that is, in its addressable market and its measurable engagement, larger than anything we underwrote in the previous cycle. The valuations being applied to companies in this category are, in absolute terms, defensible against several decades of forward execution. The problem is that defensibility against decades of forward execution is not the same as appropriateness for current pricing. A company that will be worth its current valuation in 2020 is a company we would buy at a discount to that valuation, not at it.
We have made this argument to founders in three of our portfolio companies this year, all of whom were considering rounds at prices we believed would not survive contact with their own operating realities. Two of them moderated their pricing; one did not, and proceeded on terms we declined to participate in. We are watching all three with interest.
On a Beijing Application That Orders Its Feed Entirely by Algorithm
A news application launched in Beijing in August 2012 has, in its first full year of operation, demonstrated that an algorithmically ordered feed can compete successfully against editorially ordered feeds at scale, in Mandarin, in a market where Western incumbents are absent. The product is significant for at least three reasons: it suggests that the next decade of consumer attention will be allocated by algorithm rather than by editor; it suggests that the Chinese mobile stack will produce category-defining consumer products without reference to Western precedent; and it suggests that the cost of building such products has fallen below the threshold at which a small team can iterate to scale before incumbents respond.
We are watching all three implications very carefully. The first will, we believe, be the most consequential for our existing portfolio; consumer companies whose competitive positioning depends on direct relationships with users will need to defend those relationships against algorithmically curated alternatives whose discovery loops are tighter. The second matters for our Asian thesis; the third matters for our framework on capital efficiency in consumer software.
We have not invested in this application. We did not have access to the round at terms that would have justified our participation; the founders had a clear preference for domestic Chinese capital, and we did not contest the preference. We may be wrong about this in retrospect. The thesis remains useful even where the participation did not occur.
On What We No Longer Believe (Three Items)
We no longer believe that the Asian and Western internet ecosystems will converge architecturally; sixteen years of evidence suggest divergence is the durable state. The 2010 letter described this; the intervening three years have only made the divergence more pronounced.
We no longer believe that consumer hardware companies are a structurally inferior category to consumer software; the iPhone era has invalidated this. The argument we used to make — that consumer hardware required physical capital, slower iteration cycles, and lower margins — no longer holds against the integration premiums that vertically integrated hardware-software companies can extract. We expect to deploy more capital into consumer hardware over the next five years than we have over the past sixteen.
We no longer believe that public-market scrutiny is reliably more rigorous than private-market scrutiny; in some categories the order has reversed. Late-stage private rounds, in 2013, are conducted with less diligence than the public-market underwriting of similar companies. The fact that this can be true is itself a reason to be cautious about the late-stage market.
On What We Believe More Deeply (Four Items)
We believe more deeply now than in 1997 in the family-fund structure as a competitive advantage; the cumulative experience of sixteen years has confirmed every implication our founding memo proposed and added several we had not anticipated.
We believe more deeply in long-tenured operator-founders as the dominant source of compound returns; the founders we backed in 1997-1999 who remain in their companies have, on average, produced returns several multiples greater than the founders who handed off operations after their first major liquidity event.
We believe more deeply in the discipline of saying no in years when saying yes is unusually rewarded; we cannot point to a single year in our history in which the cost of saying no, measured at the time, did not appear excessive in retrospect.
We believe more deeply in the importance of being present, physically, in the markets where consequential companies are being founded; the offices we have opened — Singapore, California, Hong Kong — have justified themselves repeatedly through deals we would have missed remotely.
A Closing Note for the Next Decade
The first sixteen years of this firm have taught us less than we expected to have learned by now. The next ten will, we think, teach us more. We intend to be paying attention.
The Partners
Winzheng Family Investment Fund · December 2013