The only question is why it took so long for venture capital firms to decide that computers and software might be able to do a better job of picking deals than mere humans.
After investing in - and hyping - Big Data as the next big thing, it seems only fair that VCs practice what they preach.
There will be the usual quibbles about whether anything can surpass the inherent genius and hard-won experience of those who have created new businesses before, but as Clemenceau once said, 'the graveyards are full of indispensable men.'
One of the challenges will be to develop algorithms sufficiently differentiated from those being employed by other VCs so that the industry does not create the same problem that hedge fund and private equity investors did: too many people with the same educational and experiential backgrounds using the same hardware and software to chase the same 'excess returns.' It is sometimes to hard to acknowledge that one is not unique or even all that different from one's competitors. But it is in sublimating the ego and identifying realistic barriers to success where real opportunity lies. JL
Greg Winterton reports in Reuters:
It looks like analytics are becoming the crystal ball of venture capital.
An increasing number of firms are turning to algorithms to comb through tons of data to dig out the deals that will result in the big exits.
In February, legendary investment banker William Hambrecht said he had hired Thomas Thurston, founder and CEO of data-crunching firm Growth Science in Portland, Ore., to find investments for his Ironstone Group, a publicly- traded holding company in San Francisco.
Now on the job, data science guru Thurston said he is using algorithms he has written himself to find businesses specializing in “disruptive innovations.”
“We’re looking for industries with big margins and that are relatively stable, and then [to] find startups that can disrupt that sector,” said Thurston, Ironstone’s CTO and fund manager.
Industries ripe for disruption include financial services, healthcare, law, real estate and transportation, especially the airline industry, he said.
But pharma is not on the list.
“Algorithms don’t work well with drugs,” he said, pointing to how unpredictable delays in FDA drug approval can affect a startup.
Ironstone is not alone. In August, data science expert Matt Oguz announced the launch of Palo Alto Venture Science, which will invest in companies using models derived from software deployed on Wall Street to rate stocks.
Founding partner Oguz said he’s raising a $200 million to $250 million fund to get the process started.
The two join two-year-old Correlation Ventures in San Diego, which is using analytics to make investment decisions for its $165 million fund.
“We’ve seen a rapid increase in using data on Wall Street, so there is no reason why venture capital can’t get more efficient by using data and some of the related tools,” said Thurston. “The days of just guessing based on how you feel doesn’t work anymore.”
Ironstone is focused on Series A deals “with the expectation that most of our investments will come in at $1 million.”
Thurston expects to make five to 10 investments by year-end, but can’t say how many he’ll do in 2014. That number will hinge on the amount of cash that Ironstone deploys in early-stage deals, he said.
“When you go to raise between $1 million and $10 million, it’s a wasteland out there,” he said, referring to the feared Series A crunch that could face seed-financed startups as they go after their first formal round of capital.
“Numerically, the early stages are the best time to invest,” Thurston said. “It’s difficult to do because of the risks. But, because we’re using algorithms, we can do deals quickly.
“The algorithms will let us do high-quality investments at scale, something VCs have never been able to do before,” he said. “Not only do yields go up, but they can scale in line with the science, so they don’t have to do large, large deals to get the money out.”
0 comments:
Post a Comment