Umair Haque has a post getting a fair amount of attention on “How to Fix Venture Capital.” It is in response to Fred Wilson’s post about needing a New Path to Liquidity.
I don’t mean to be too annoying to our esteemed colleagues in VC, but isn’t it just like a VC to worry about the endgame rather than the beginning? I suppose they would say of me that it’s just like an entrepreneur to look at the beginning rather than the endgame. Let’s look at the two, but I think the beginning is the more important phase than staying the course later and Google provides a perfect example.
The VC liquidity complaint is that you can’t take a company public and the M&A game is no great shakes either. Why can’t you take a company public? Two reasons. First, the economy has closed the IPO window for the time being. Underwriters like it to be frothy so the stock goes up from wherever they price it. It also helps if the newly public company can make its numbers in a good economy. The window has not really been open very wide for some time, but there were at least some IPO’s in recent years (I was fortunate to participate in one such at Callidus). The IPO Window is a temporary problem though, and can’t be viewed as symptomatic of the whole industry.
The second problem is that companies have to be a lot larger, a lot more predictable, and a lot more profitable to go public. Figure circa $100M per year, very profitable (or growing like a Google in its heyday), and there needs to be reason to believe the first public year’s numbers are almost a mathematical certainty.
For some types of companies, the mathematical certainty thing is extremely hard. You either need a tremendous number of small transactions (the Web 2.0 model) or you need a SaaS model that will average your transactions over a broader window than a quarter. For others, profitability is problematic.
But there is another problem here. That $100M revenue number is a very high bar for a small company. It means they have to run for a lot more years. In the old model, it worked something like this:
- Build a product first year.
- Get to $1M and fine tune the second year.
- Double every year after until you go public or get bought.
Note that to reach $100M, the progression is 1,2,4, 6, 16, 32, 64, 128. Figure on 6 or 7 years of selling after 1 year of building and you have 7 or 8 years. That’s too long for an industry (VC) built on 5 to 6 years until liquidity. The extra years can drag the firm’s IRR down and make it less attractive versus just buying the old S&P500 index fund for the limiteds.
It gets worse: not every company is a $100M a year business. Umair says the biggest issue is companies selling too soon because they were pushed by VC’s to do so. Given my IRR remarks, maybe. Umair lists Myspace, Skype, Last.fm, del.icio.us, and Right Media as examples of companies that sold too soon. I have to bring up Pointcast in counterpoint: a deal that didn’t get sold soon enough. They didn’t take a News Corp offer for $450 million and were bankrupt 2 years later.
Deal or no deal? What of Umair’s examples?
Well, MySpace reported $10M in profit on $550M for fiscal 2007. It was founded in 2003, so it definitely blew past the growth hockey stick I gave above. How did it compare to Google? It was incorporated in 1998 and they raised a little over $1M for their initial capital, and they started selling ads in 2000 and their S1 was filed in 2004. FWIW, Google’s progression was $86M in 2001, $347M in 2002, and $961M in 2003. But, the profit went from $11M to $186M to $342M during the same period.
Therein lies the rub: Google was ridiculously profitable from the beginning. MySpace is just getting to Google’s earliest profit level as they are 5x larger. That tells us it would take ridiculously more capital to get there with MySpace than a Google. Even today it is unclear for many whether Social Networks have really discovered their business model and hence ultimately what their value might be. All companies are not Googles, and no amount of stumping about resolve to stay the course and not sell out will change that. When you’re bleeding capital out every pore and your eyeballs, you sell.
Why was Google so much more successful? The answer to that lies in the beginning, and not in the endgame, for it is clear that Google was extremely successful right from the start. When you’re tripling every year and wildly profitable, it’s easier to sit back and let it ride, or at least, as Paul Graham puts it, to have the balls to ask for an extremely high valuation. But how do you get into that position to start with?
Let’s ask another question: Is Google just a total fluke and not repeatable? One prominent VC said to me, “We’ve built our business on the idea that we can have a Google in every fund but it seems clear the entire industry gets one Google per decade.” He went on to say that many of his colleagues felt the answer was to wait longer before getting involved. If Google was so successful so early, maybe the answer is to conserve cash and wait for small companies to demonstrate traction before investing. Apparently they were almost flirting with the LBO end of the spectrum.
Hence we have a model today where “pro” VC’s largely don’t invest in a guy and slideshow, or even perhaps a kid and class project as the Google guys had. They want product and some customers. Let the Angels worry about investing before that stage. I’ve written before about how the decline of VC fund returns tracks pretty closely their decision not to play in seed funding. Is there a cause and effect, or is it all just a function of how the economy is behaving? Returns are still under a lot of pressure, so it behooves us to figure it out.
Something else to think about: Google had significant technology, and hence significant competitive advantage and barriers to entry. After their initial $1M’ish investment, which got the original product together, Google followed up just a year later with John Doerr at Kleiner and Michael Moritz at Sequoia splitting a $25M equity financing for Google. That’s quite an early round of investment and it put Google in the driver’s seat in terms of pressing home the advantage of their technology as rapidly as possible. Note that this investment came before the company had any revenue whatsoever–they didn’t start seeling ads until 2000!
In 1998, when Google got their first $1M, VC’s were making twice as many seed investments in software as they do today. See my post for the data. Note that even in 1998, seed investments were half what they had been and Google started out with Ram Shriram’s money before they could get to KPCB and Sequoia.
Google raised $25M before they had a dime in revenue at a time when they were the 5th or 6th search engine into the game.
Although Google seemed “me-too” as a late comer to the Search game, in fact they were highly differentiated. It looked different from the start to users, and it had some radical technology innovation behind the scenes.
Entrepreneurs today have to go Angel. I read recently that the average Angel investment is $250K. Certainly the sweet spot is under a million and probably $500K or less. Given that Google spent $1M to get to their same place in the world, what does having half as much capital in a later (more expensive) time do to the startup world? It means you’re going to build something much easier than Google. There is even some thought of trying to get results for 1/10 that in the Y Combinator model. What can you build for $25K? What if the market is actually efficient enough to figure out that means it isn’t worth as much? DOH! I hate when that happens!
Worse, Angels are often much higher maintenance than professional VC’s. This anecdote from Stowe Boyd is not atypical among entrepreneurs I talk to:
“Today I had lunch with a smart, seasoned entrepreneur who told me about a 4-inch deal binder he had been forced to create for angel diligence. As I said to him: Run. Hide. Any angel who wants that much security in an early-stage deal is to be avoided like a banker.” Ugh.
Picture how much eye-off-ball time is needed when you’re preparing 4-inch binders to raise $256K every 6 months to keep your deal going. Compare and contrast that to raising 18 months of runway in the old model. A good friend who helps entrepreneurs raise money and is herself a former VC said her husand had to call on 60 different investors before he got funded. In the old days 12 or 15 meant you had the wrong idea and should move on. Not hard to guess who got more done that mattered to the business: it was the entrepreneurs in the older time.
With so little capital to build a product, we see endless simple me-too plays. It’s not that hard to build a Web 2.0 property. Yay! It’s not that hard to rip-off a Web 2.0 property. Boo! Did you see how easily Google copied 37Signals with Huddletalk? Get ready for a big steaming cup of FAIL if you think building trivial products for no money down is going to be a winning model. You’re going to pay for it in the noise and confusion when 10 others copy your idea or variations of it in no time flat. Fast Follower strategy is now going to be Fast Pillaging Horde on your Trail. Yikes!
So we get a lot of vitamins, because it costs too much to build painkillers. Software for business, where an ROI is easier to measure than with Twitter, is almost by definition going to cost 2-3x more to build. The IT gatekeepers will simply insist that more of the basics are covered before they’ll look at it.
How To Fix Venture Capital?
Umair writes that it’s just a matter of conviction. As he says, “rediscovering the purpose to put true, durable, meaningful conviction behind investments.” Aw, come on, though. Given Google’s numbers, conviction was easy once they got rolling. Profitable and $80M in 2001? Tripling every year and even more wildly profitable? Sure we’ll stay in with those results. I’m with Paul Graham in saying that Umair is wrong, although I love his passion.
The real conviction came at the beginning, when Doerr and Maritz put $25M to work before a dime of revenue was visible, and when Sriram gave some kids $1m to build a prototype. Real conviction is building out a big expensive technology vision. That’s right, a technology vision. Google’s 2001 financials tell an interesting story. In 2001, they spent as follows:
Cost of Revenues (Data Center Ops): 16% of revenue
Research and Development: 19% of revenue
Sales and Marketing: 23% of revenue
General and Administrative: 14% of revenue
When was the last time you saw a company with $86M in revenue spending so much on R&D? For comparison, here are Salesforce.com’s numbers when they were doing $94M in revenue, a similar size:
Cost of Revenues: 15% (SaaS can be delivered at a similar cost to Search, it seems)
R&D: 10% (half what Google was spending and more what the world thinks is in line)
Sales & Marketing: 60% (clearly hugely inefficient, hence the whole “Sales 2.0″ and Freemium movements)
G&A: 22% (also much less efficient than Google. Perhaps this is a function of SOX.)
What if the problem today is that there is not enough capital available to invest in difficult technology? What if you need to build a product that will cost $2M to get to an interesting level of traction? Worse, what if it costs $5M? What if you need $25M like Google did? You simply can’t get there on Angel financing and the VC’s aren’t there either. So the deals never get started.
Think of it this way: If you had a chance to spend 2x on innovation versus 2x on sales and marketing versus the competition, which one do you think provides the more lasting advantage? Which one is strategy and which one is merely tactics? If you’re a VC, which one do you prefer to spend your capital on and does your current investment strategy really reflect that thinking?
This is not an excuse to ignore profitability, BTW. We started out looking at how unprofitable MySpace is compared to Google. Part of finding your product/market fit is profitability. Too much emphasis on growth without regard to profit may be questionable. People are fond of talking about how far Google got before a profit, but it doesn’t look that way to me at all compared to MySpace.
To fix VC, start at the beginning. With portfolios, you want some counter-correlation anyway. If everyone follows the same model, you are prone to the same risks. Put seed back on the table. Look for those big ideas that are worth a more expensive seed. Revel in the idea that not many are getting funded that way so perhaps there will be little competition. That lowers your cost to sell dramatically when there is less noise. It also speeds adoption when customers find out you have the only game in town. At the same time, consider Paul Graham’s advice. Spread the net with smaller investments to cover more bases until you get the breakout. It seems to me that VC’s today are stuck in the unhappy middle ground. If they’re all there, their deals are all highly correlated, and they going to regress t the mean. Not a happy place.
Fred, you’re right, VC’s need to reinvent the game. But consider the beginning as much as the end.
You just might find another Google.
Prototype Invest is a new firm R/W Web is talking about whose mission is to take equity and create the initial prototype for startups. Sounds like others are having a problem finding seed money to build their prototype.