Over the last 13 years, Stan and I have found ourselves drawn to focus on businesses that have network effects.  It started intuitively, then became explicit in 2003. Since then, we’ve started or advised over 25 companies whose core business was either 1) building a network of people who wanted to communicate, or 2) a network of buyers and sellers who wanted to transact.  Some of these companies have both.

There are others who have admitted a similar affection for these businesses, including David Sze and Reid Hoffman of Greylock, Fred Wilson of Union Square, Bill Gurley and Matt Cohler of Benchmark, Jeremy Levine of Bessemer.  And no wonder.  If you look at the biggest tech returns of the last 15 years, many of them were either marketplaces or networks, and if you consider companies that became worth more than $10B, nearly all of them fall in this category.  We’ve done a detailed analysis of this ourselves, and recently, James Slavet, also of Greylock, published similar findings.

So why isn’t everyone focused on these?  Increasingly, they are, of course.  But the fact remains, they’re f**king hard to pull off.  There’s a lot of art/luck in them, and that’s hard to predict and thus hard to invest in.  Over the years, we’ve watched/stumbled into/invented many tactics to manufacture a two sided network, to A/B test your way to virility, to foster liquidity and tipping points, to buy your network inexpensively, to iterate until you find the right subject matter or value proposition, build the right retention and feedback loops, etc.  These lessons can be applied today to increase the chance of successfully creating a functioning marketplace or network, or both.

The first step is even realizing you are attempting to develop a business that fits in this cohort of companies, and that there are now lessons to learn from prior successes.