Smashing the system
Spotlight on: the OpenGamma Platform
finance industry has, for the past several decades, relied on
expensive in-house and proprietary software, unaware of the open
source revolution taking place elsewhere. That is, until
London-based startup OpenGamma ripped up the rulebook
with their freely-available platform.
Yet despite having been nominated for a JAX Innovation Award earlier this year and securing $15m in funding in August, the company still remain unknown to most tech commentators.
Perhaps it’s because the ideas driving OpenGamma are so complex and niche. Despite years of talking to the press, CEO Kirk Wylie still struggles to explain the idea behind the OpenGamma platform. If asked at a cocktail party, how would he describe his company’s software?
“I say, 'We make software for hedge funds and investment banks,' and I leave it at that,” laughs Wylie.
Also present are the company’s Java guru Stephen Colebourne (also co-lead developer of JSR-310) and Chief Technology Officer Jim Moores, who provide a little more detail. “The primary idea of our software - well, there’s a couple of things,” says Moores. “One is, if you're a trader, it allows you to effectively try out different strategies and see what the results of those might be and what effect that may have on your profit.
“Then the other aspect of it is more of a risk control - so effectively giving you an idea of statistically, what's the most amount of money you're likely to lose on a statistical basis, if the market shifts in certain way. So, let's say, the fed announces they're putting up interest rates by 1%, then you can have already simulated that, and estimate what effect that might have on your portfolio.”
What differentiates OpenGamma from the competition is its open source nature. While this concept might appear to be an obvious one, the niche that OpenGamma hopes to reach has relied for the past decade on proprietary or in-house software.
“This is a market that 47% of the industry is still built everything in-house,” says Wylie. “Software vendors are only 24%. I think that's a sign that vendors have failed the market.”
OpenGamma isn’t a framework, or product. Instead it’s a platform upon which companies can then build upon. By providing a reliable basis to base their internal systems on, OpenGamma hope to save their customers time and money. As Colebourne puts it: “we're writing the boring bit of code that they all write”.
If the same logic of the world of finance was applied to web servers, says Wylie, “every single company that wanted to host a website [would have] to write Apache from scratch, or buy the Netscape server”.
OpenGamma is “shining a huge, great open-source light on an entirely different business model”, says Colebourne, a revolution that will be “jolting everyone else awake”.
Utilising open source software is of course not only cheaper - it provides transparency too, an essential feature in an industry where poorly-coded software can cost millions.
"The open source part of that is going to become increasingly important,” says Moores. “Because at the moment there's very little transparency. None of our competitors have any transparency at all in the way that they determine those numbers.”
“You'll have a vendor that will have a document that will explain the mathematics of what's going on,” explains Wylie, “but that's the closest they'll come. So, do you know they're implementing that mathematics correctly? No. You kind of trust that they are, but you're never really sure. So, a lot of firms, to get to the kind of transparency that they want, have traditionally built from scratch.”
“This black box world is not optimal for anyone, and I think if you looked and said, 'is it right that there's these trillions of dollars of notional, of contract, that's being traded every day, and no regulator has ever seen the source code for any of the code that does risk on that in any way?' You'd sit there and say, 'No, that's kind of crazy and scary'.”
The idea for the OpenGamma platform originates from several
years back when Wylie was working on in-house software similar in
“I was in New York for a project kickoff meeting, for a project that architecturally looks not dissimilar to OpenGamma today,” he reminisces. “It was 3am, [I was] jetlagged, and it dawned on me that we were about to spend a minimum of ten man-years to build something that delivered absolutely no proprietary value to the company whatsoever.
“We may as well have flushed that money down the toilet, because all this was, was plumbing. It was just the stuff you have to do - everyone was doing it.”
Wylie, a self-confessed “complete Silicon Valley, pain-in-the-ass prima donna programmer”, had previously met Moores and third co-founder Elaine McLeod met while working on internal finance software in Smartspread. It wasn’t until the three found their services “no longer required” by their employers in the aftermath of the 2007 crash that they begun seriously considering the business model that would eventually become OpenGamma.
“We spent about eight months doing nothing but architecture and business model and market research and lining up potential customers and dealing with funding and all that,” says Wylie. “So the official start of August 2009 had eight months of pretty solid business plan work behind it before we even got there.”
One choice made during this planning period was to code the majority of the OpenGamma platform in Java.
Why Java? “We get that a lot,” says Wylie. “Jim?”
“It's a combination of factors, really,” responds Moores. “It's a robust, stable, modern language. It doesn't have all the cutting-edge features that some people might want, but I think it's generally accepted in the industry. We had to make a decision, really, about - were we going to go with something really cutting-edge, and choose, I dunno - Scala or Ruby -”
“Go!” shouts Wylie.
“Something crazy like that,” continues Moores. “The reason we didn't go for that end of the scale really was, y’know, there's not the maturity there. Customers would get scared off by it.”
“One of the advantages of open source,” adds Wylie, “is that people can read, and understand, and parse and change the source. And if we do it in a language that our target userbase isn't allowed to use at work, then what's the benefit?”
interview with The Kernel earlier this year, Wylie was quoted
as being “annoyed” that OpenGamma is “ignored” by most tech
commentators, despite raising $23 million in funding already. He
says he was misquoted, and that his annoyance stems not from
OpenGamma being ignored specifically, but a relentless focus on a
specific brand of technology startup.
“I think the point is that it misrepresents the tech industry in London,” agrees Moores.
“The relentless focus on consumer internet stuff sends the message to people that could be entrepreneurial that that's the only way to go,” says Wylie, giving the example of software for aerospace companies, a niche he sees as underexploited. “What I don’t those guys to say is, ‘if I want that entrepreneurial lifestyle, if I don’t want to work for the aerospace companies, I have to give up my domain expertise’.”
“By having all this press attention on the one space, government policy might then follow that precedent,” adds Colebourne, “but actually it should be following the far more bigger-employing, bigger-impact enterprise-based stuff.”
Still, ignorance of OpenGamma by the wider press doesn’t affect the company’s bottom line. “Would an article with my photo on page 14 of the Daily Mail result in a flood of people calling in saying, ‘I would like a quarter-million dollar-a-year support contract on your open source derivative trading software? No’,” jokes Wylie. “We get the attention of the people we need to get attention of.”