The Math Behind Venture Capital, and the Need for Home Runs

Recently, I discussed how venture capital in the tech sectors is weird: every investor is swinging for the fences. Rachel Chalmers describes the financial incentives (boldface mine):

Besides being custodians of the fund and therefore honor-bound to seek the best outcomes, investment managers are compensated in part by carried interest. Carried interest is a cut of returns to the fund after (a) all capital has been returned and (b) the fund has met an agreed-upon rate of return. Let’s say the VC firm has a $150m fund and a preferred return to investors of 3x, or $450m. The partners in the VC firm don’t even start earning their carry check until that $450m threshold has been reached.

What does this mean for you, everywoman entrepreneur? Let’s assume this firm has offered to invest $15m in your company in exchange for 15% of equity. At that rate, said VC firm can make ten such investments. We all know that nine out of ten startups fail, so the VC is betting that one of the ten of you can return the fund, and then some.

What do you have to do to make the VC’s 15% of equity worth in excess of $450m? We are now talking about a $3bn outcome.

Whatever it is you have to do to earn this – take the company public, somehow impress Mark Zuckerberg – your investor/co-owner is now focused on persuading you to do. That’s why they’re starting to ask if you would consider moving aside to let an experienced public-company CEO take over.

By the way, the odds of achieving a $3bn outcome are so far from being in your favour, it’s absurd. Not winning-the-lottery absurd, but an order of magnitude worse than that old one-in-ten rule of thumb. Let’s leave aside the outliers like WhatsApp and focus on enterprise software, a business I know and one that’s slightly less subject to the whims of wanton Gods than consumer-land. As an industry analyst, I covered 1054 enterprise software companies over a 13-year span. Of these, 75 were incumbents or enterprise end users of other vendors’ software, so let’s say 1000-odd were venture-backed startups. Where did they all end up?

Well, 188 of them were acquired, in deals that ranged from the sublime (Citrix paid $500m for XenSource, which had trailing 12-month revenues in the $1m range) to the ridiculous (many, many face-saving exits of failed firms consisting of a nominal fee in exchange for their patent portfolio.) Note that an exit in the tens of millions, which is a champagne-and-confetti Retirement Level Event for the owners of a bootstrapped company (like Sleepycat Software or Platform Computing), can be a disaster for a venture-backed firm. In particular, since the investors own preferred shares (which means they get paid first) and the employees own common stock (which means they only get paid if there’s anything left over), the people who actually did the work can walk away from a multi-million dollar acquisition with nothing to show for it but sadder and wiser hearts.

Back to my cast of one thousand: 28 of them failed so hard they don’t even fog a mirror any more. Google their corporate site and get domain-squatter linkspam. No face-saving exits for them. I won’t name names. Eight went public: eight out of a thousand. Their names are Opsware, Rackspace,, ServiceNow, SolarWinds, Splunk, VMware and WorkDay.

What this means is that you are more than three times as likely to crash your startup as you are to ring the Nasdaq opening bell. Your chance of utter failure is around 3%.

Your chance of going IPO is just under 1%.

The key part, other than the dismal success rate (you would be better off opening a restaurant which isn’t exactly a low-risk proposition), is that many companies that would be successful turn out to be failures–at least for the talent–simply because of the hunger of the VC firms. There has to be a better way to mobilize capital than this.

Posted in Bidness, Internet | Leave a comment

Links 4/17/14

Links for you. Science:

One drug to rule them all: Researchers find treatment that kills every kind of cancer tumor
Dear Jenny McCarthy
Large Samples: Too Much of a Good Thing?
GOP pushes funding cuts for social science work


The ‘Real Racists’ Have Always Worn Suits
Gefilte Fish: Why, Oy Why?
The History Of Manischewitz; Or, Where Did That Sweet Grape Wine Come From?
Are Americans really jingoistic yahoos? A cautionary statistical tale
Three Expensive Milliseconds
The New Age: Leaving Behind Everything, Or Nothing At All
Confirmed: “Men’s Rights Activism” Is For Misogynists Without God
How Burrowing Owls Lead To Vomiting Anarchists (Or SF’s Housing Crisis Explained)
How Being a Doctor Became the Most Miserable Profession
The Meaning of Oaths and a Forgotten Man
The big losers in NYC charter fight: students with disabilities

Posted in Lotsa Links | Leave a comment

Putting the Bundy Ranch Dispute in Context

A slightly different perspective on whose rights are being violated during the whole Bundy Ranch dispute:

Cliven Bundy’s family worked their ranch land since 1877. The family claims ancestral and sovereign rights.

On Monday, the I-Team received a map from the Moapa band of Paiute Indians showing how the land the Bundy ranch is [on] was promised to them by federal treaty.

That is until federal troops forced the tribe out and families, including the Bundy family settled in.

I hope the Paiute take Bundy to court.

Posted in Uncategorized | 3 Comments

Some Thoughts on “Rescusing US Biomedical Research From Its Systemic Flaws”

By way of the intertoobz, we come across this open access PNAS article about the problems facing biomedical by Bruce Alberts, Marc W. Kirschner, Shirley Tilghman, and Harold Varmus, big names all. Before I get to the flaws, it’s worth noting that this article does something very important: it puts the lie to the notion that we have a STEM shortage, at least in the biomedical sciences. Admittedly, that’s nothing a gajillion science bloggers, including yours truly, haven’t discussed, but Alberts et alia are Very Serious People, so maybe this reality will get some traction. So the first half of the article, which lays out the problems, is excellent*.

Then we get to the solutions. Before I get to the individual items, the immediate problem that the article skirts around is that there are too many academic PhDs chasing too little funding. A culling is occurring and will continue to occur; the question is what are the criteria that will be used for the cull. I share DrugMonkey‘s concerns that there is an air of elitism surrounding how those decisions will be made. The other ‘contextual’ issue is that there is an assumption that NIH research should still mostly be the purview of academic principal investigators (PIs)–this is an assumption that needs to be challenged.

Onto the specifics (not all of them).

I absolutely agree with their stance on PhD training grants–a point I’ve made before. Institutions should only receive NIH funding for graduate students if they make a compelling argument that they have a good training program. If you want to hire a lab assistant (or a ‘junior’ postdoc), then pay lab assistant/postdoc wages and benefits. Students should have a comprehensive training program that focuses on students**.

And then there’s the staff scientist section. This is where thinking about science solely occurring via the Solitary Heroic PI and Her Band of Plucky Followers goes off the rails. From the article (boldface mine):

We believe that staff scientists can and should play increasingly important roles in the biomedical workforce. Within individual laboratories, they can oversee the day-to-day work of the laboratory, taking on some of the administrative burdens that now tend to fall on the shoulders of the laboratory head; orient and train new members of the laboratory; manage large equipment and common facilities; and perform scientific projects independently or in collaboration with other members of the group. Within institutions, they can serve as leaders and technical experts in core laboratories serving multiple investigators and even multiple institutions.

We recommend increasing the ratio of permanent staff positions to trainee positions, both in individual laboratories and in core facilities that serve multiple laboratories. To succeed, universities will need employment policies that provide these individuals with attractive career paths, short of guaranteed employment. Also, granting agencies will need to recognize the value of longer-serving laboratory members. If adopted, this change would help to bring the system closer to equilibrium. There is precedent for such a policy in the intramural NIH research program, which employs many well-trained MSc and PhD graduates as staff scientists to conduct research.

Having been a staff scientist, there are a few problems here. First, core and within-lab positions need some kind of backstopping. If after five years, the money for the position vanishes, then you’re screwed. This might work for science spouses, but I can’t see the R01 mechanism working, except for the funding elite. The other thing is that this is very geographically-dependent. If you’re in Boston (or a few other locations), when the funding dries up, there are other options locally. But if you’re at a small university–or a large one not near anything else–when the job disappears, you’re not just switching jobs, but where you live. With all the pissing and moaning about Boston and other elite coastal cities, I imagine this would be viewed as a problem. The NIH campus has options for staff scientists who need a new position, so I don’t know how broadly applicable, outside of San Francisco, Boston, and possibly Research Triangle, that model is unless NIH changes how it funds science.

Which brings us to this grant making suggestion:

Inertia and financial dependency favor continuing large research programs, so sunset provisions should be built into all new programs and orchestrated team efforts. To combat the tendency for fields to become parochial, agencies should develop funding mechanisms that encourage the growth of new fields, both by direct support for new science and by a rigorous regular evaluation of existing programs.

This sounds great, but, in practice, what often happens is just as a large research group is getting good at what it does, the funding is slashed or disappears and all of that expertise vanishes. Obviously, some programs have an obvious end–sequencing the human genome. But if you want stability for staff scientists, then long-term, large scale research programs are necessary (or marry wisely!).

Finally, the grant review proposals are just daffy. I would argue we need more junior people closer to the research as well as non-academic researchers (i.e., from institutes and companies when appropriate) as reviewers. Should only an older and elite subset of the eight percent of PhDs who get tenure be the only ones making funding decisions? I have doubts.

So what to make of this? The problem is defined superbly, but I think there’s a lot of bias towards elite academic institutions and the single PI model. Unfortunately, there are a lot of assumptions in the article about how science should be done, who gets to do it, and who decides who gets to do it, none of which are explicitly stated (other than senior people are wick-ED smaht!). If we want to solve the funding problems, we need to bluntly and honestly lay those assumptions bare.

*One key point is that overheads are now including the costs of building new facilities to house scientists; maybe not the best use of overhead dollars….

**The only pedagogical classroom training I received consisted of “when you erase the blackboard/whiteboard, do it from top to bottom, not side to side, because side to side makes your ass wiggle.” Not kidding.

Posted in Funding, NIH | 1 Comment

Links 4/16/14

Links for you. Science:

Amber Waves of Woo
Fecal transplant most cost-effective first-line treatment for recurrent CDI (headline isn’t entirely accurate; still worth reading)
This Fish Crawled Out of the Water…and Into Creationists’ Nightmares
The Mystery of the Glow-in-the-Dark Civil War Soldiers


Let them eat McMansions! The 1 percent, income inequality, and new-fashioned American excess: We have sprawl, wars over cheap gas, stagnant wages and longer hours because your boss wants this awful, ugly house (excellent)
One Way to Get a Safe Abortion in Brazil: Pretend You’ve Been Raped
Sue Townsend: how the welfare state left me and my kids scouring the streets for pennies
Moar homeopathy in Davis
Angry White Men W/Guns Threaten Violence And Win. Good Luck To Black People Trying The Same Thing
Wait Hold on Treason
Wrinkled Ghosts
David Berliner on PISA and Poverty
The Credibility Gap

Posted in Lotsa Links | Leave a comment

Does the Far Right Understand They’re Not the Only Ones Who Can Buy Guns?

What I don’t understand about the far right’s embrace of Cliven Bundy, who has been alluding to a violent response to any attempt by the federal government to get him to pay over $1 million in fees he owes, is that they believe they’re the only ones who can get guns. Before I get to that, Steve Benen summarizes the problem:

…it’s unsustainable to think a group of well-armed extremists can simply block the enforcement of American laws in the United States. It’s perfectly understandable that the Bureau of Land Management saw a crisis unfolding and pulled back to prevent bloodshed, but there’s an obvious problem with establishing a radical precedent: you, too, can ignore the law and disregard court rulings you don’t like, just so long as you have well-armed friends pointing guns at Americans.

To put it mildly, that’s not how the American system works. Indeed, that’s not how any system of government can ever work.

Maybe the mistake Occupy made–or workers who think their unions were unfairly busted–was to not bring guns. Worse, the one group of people who might be able to calm them down has no gain in doing so.

Posted in Conservatives, Fucking Morons | Leave a comment

The Heartbleed Hack and Misunderstanding the State of Our Tech Infrastructure

The state is bad. Before we get to the Heartbleed security hack, which was exploited by the NSA, it’s worth revisiting our usual plaint regarding the state of science infrastructure:

One of the ridiculous things about many depictions of science in TV and movies is the notion that there’s this huge infrastructure: shiny labs (which are always neat and spacious), high-tech this and that, and an army of workers to solve a problem. The reality is that much of our scientific knowledge in any subdiscipline is held by a few people who are operating on shoestring budgets with inadequate resources. To put it bluntly, we often lose considerable knowledge and materials when an older faculty member or researcher dies or retires (in my own subdiscipline of microbiology, there are several valuable collections that would be lost if a single freezer broke for an extended length of time).

So we were saddened but not surprised to read this about the Heartbleed hack (boldface mine):

What’s amazing, however, is that the code that contained this bug was written by a team of four coders that has only one person contributing to it full-time. And yet Henson’s situation isn’t an unusual one. It points to a much larger problem with the design of the internet. Some of its most important pieces are controlled by just a handful of people, many of whom aren’t paid well — or aren’t paid at all….

The sad truth is that open source software — which underpins vast swathes of the net — has a serious sustainability problem. While well-known projects such as Linux, Mozilla, and the Apache web server enjoy hundreds of millions of dollars of funding, there are many other important projects that just don’t have the necessary money — or people — behind them. Mozilla, maker of the Firefox browser, reported revenues of more than $300 million in 2012. But the OpenSSL Software Foundation, which raises money for the project’s software development, has never raised more than $1 million in a year; its developers have never all been in the same room….

In some ways, there’s a bug in the open source ecosystem. Projects start when developers need to fix a particular problem, and when they open source their solution, it’s instantly available to everyone. If the problem they address is common, the software can become wildly popular in a flash — whether there is someone in place to maintain the project or not.

But our awesome economic system is doing an excellent job of mobilizing and allocating resources. Or something.

Perhaps there’s a cost to high frequency trading after all?

Posted in CIA Spy Shit, Funding, Internet | 1 Comment