It is fascinating to see how people are using open source. I'm part of the "old guard" of open source, I suppose, delivering an open-source alternative to a tired market ripe for commoditization and innovation. But other companies, like OpenAds (open-source advertising server), Path Intelligence (tracking shopper flow based on the open-source GNU radio), Chumby (open-source consumer electronics/hardware), etc. are taking open source into new markets.
Today, I was fortunate to meet one of the most interesting open-source companies I've seen in a long, long time: Marketcetera. Marketcetera provides an open-source trading platform that hedge funds and others use to process and deliver trades through a brokerage to an exchange (like NASDAQ). It's like proprietary, expensive FlexTrade, only not proprietary...or expensive.
The market for this kind of platform is not huge today, as the founders, Toli Kuznets and Graham Miller, told me today (roughly $500 million for custom development, but probably not including packaged software like FlexTrade). But with more and more trading moving from people to algorithmic processes (30-40% in the US today, jumping to 50-60% by the end of the decade), the market will grow accordingly.
Besides, I can think of a range of other uses for this sort of technology beyond hedge funds.
As D.E. Shaw and others demonstrate, algorithmic trading is the future. Marketcetera steps into this opportunity with an open-source trading platform that delivers fast, flexible and reliable securities trading tools for financial services professionals, with the added benefit that it is well-suited to "Not Invented Here" financial services companies. These are the sort of companies that have more developers than most software companies - they build most of their software, not buy it. Giving them an open-source platform allows them to build their own, but from the foundation provided by Marketcetera.
So, Marketcetera provides the base platform, and then offers value-adding services on top. Initially, this includes support and maintenance. (Most hedge funds are under $100M and so maybe they have a few IT people that are creating pricing logic and writing scripts, but they don't have a big staff to build a trading platform.)
But the real value of an open-source model for this market is the availability of a couple of other, potentially large revenue streams:
- Market data (globally an $11B industry). Stuff like this. Things like bid/ask quotes on given exchanges.
Surely, I probed, this sort of data is commoditized and already better supplied by Reuters, Bloomberg, and others? Yes, replied Toli, but the value of market data is influenced heavily by speed (Reuters charges a heavy premium for its market data because, as it says, its data comes at the lowest latency/fastest speed possible) and reliability (many traders go with multiple data vendors to ensure they're getting good data).
Marketcetera's initial pitch would be that its market data integrates best with its platform. In other words, if a company is using its platform, Marketcetera will be the best source of data for that platform, not because the data is different from what a Reuters could provide separately, but precisely because its market data is not separate from the trading platform. Integration with the platform trumps best-of-breed, separate components, is the thought.
- Brokerage services (Hedge funds and others may build their own proprietary trading systems, but they still need a broker to process trades with an exchange, which is another $11-12B industry). Marketcetera will initially resell other firms' brokerage services, but ultimately it could get regulatory approval and be certified as a broker and take a commission on every trade through its system.
Sound interesting? Absolutely. This strikes me as a very interesting way to combine open source with Tim O'Reilly's ideas on data as the true value driver going forward for "software" companies. Marketcetera is not in the software business. It's in the services business, with an open-source software platform making the services possible.
By the way, in case you're like me (i.e., not an expert and hardly even a novice in financial trading systems), here's how it works:
In the case of a hedge fund, it decides which stock (or currency, or whatever) to buy or sell, which it then sends as an order to a broker. The broker then forwards the order to an exchange, which settles the order and transfers money back to the hedge fund.
I asked, what value does your system provide to the hedge fund? Presumably they can already issue orders.... Toli and Graham responded that their system is a way for hedge funds to send a lot of orders at the same time. The customers may have proprietary systems that help them algorithmically determine what to buy, then they could connect that system to Marketcetera's to format, process, and execute the trades to the brokers in a way that the broker's system can understand and automatically route to the exchange. The trading platform then collects info from the exchange on the return trip and provides basic risk and position analysis for the hedge fund.
The company is still in the development stage of its product, but has a few paying customers now and expects to be in production with these customers by the end of August. Marketcetera has customers using the platform for currency trading as well as equity/options trading. Others are using it without a relationship with the company, just as they should in an open-source community.
If you're interested in taking a look at what Marketcetera has built, you can download it from the company's site. You can also get it as a virtual appliance. Marketcetera's software is 100% open source and is licensed under the GPL.
Days like this cause me to remember that we are in open source's infancy, not its old age. Things are great now, but they're only going to get better with projects like Marketcetera's.