Turning data into money at speed: The Capitalogix story
Decades ago, many people bought and held stock for years. But the world has changed. The average holding time for a security shrunk from years to months to weeks to days to hours. Today it’s less than 20 seconds. That’s why Capitalogix CEO Howard Getson said, “if you’re not using real-time data, you’re playing a totally different game.” Real-time insight is essential for Capitalogix—and IBM delivers the data platform that helps them achieve it.
The need for real-time insight
On the surface, Capitalogix may look like a typical hedge fund, but it’s really a data science firm in disguise. Their data science professionals build models that analyze capital markets and economic conditions all over the world, trying to identify opportunities for investment. Then their trading division acts on those opportunities. This is essentially a never-ending battle within a global marketplace, won through advanced analytics.
Capitalogix can’t ever know too much or move too quickly, and any advantage they gain can be fleeting. So their priorities are for the highest-possible speed of analysis on the broadest-possible range of data. According to Chris Jordan, CEO of their technology partner iOLAP, Capitalogix tries “to analyze everything that’s going on in every market everywhere. The platform has to be able to support that.”
This demand for a platform with massively parallel processing and ever faster data analysis has driven the evolution of their IT systems. Over the years, Capitalogix has moved from Microsoft SQL Server to Netezza and finally to the new IBM Integrated Analytics System (IAS) platform.
The rationale at Capitalogix was simple: leverage a data environment that helps their data scientists develop custom algorithms, then apply those algorithms at maximum performance on the maximum range of data. They can’t focus on just a few variables. Their focus neesd to be everywhere, all the time, on nearly everything.
Value from a variety of data
When a firm wants to focus on nearly everything, the data challenge goes far beyond ticker outputs and nice, clean spreadsheets of financial results. Capitalogix is committed to finding the value in all data. Their CEO explains it well: “There are lots of different ways of figuring out what’s happening in the economy,” using inputs that could include:
- Transactional and time series data
- Unstructured data and sentiment feeds from social media
- Satellite imagery of shipping vessels that reflect market demand
- Aggregated data on when credit cards are declined
That’s why Capitalogix needs a platform that lets them “take it in, curate it, refine it, figure out how to use it, reformat it, transform it—and ultimately make it part of a unique competitive, sustainable, competitive advantage.”
This kind of far-ranging synthesis of different data sources requires unique data algorithms, which are Capitalogix’s specialty. Rather than having one trading tool that they try to apply wherever the market allows, they recognize new situations, then create trading tools no one else has used before. As Derek Ainsworth, data science product manager, said, “a lot of trading firms, they look at the market and they say, ‘I have two tools. And how many places can I use that tool?’ We develop tools.”
This focus on developing their own models is another reason why a robust, in-house data platform is essential for Capitalogix. Combined with their need for maximum speed, the case for an appliance was strong.
The case for an appliance
Capitalogix chose an in-house data appliance because of their combined need for performance, control and economics. Simply put, when the computing platform was being used at maximum speed around-the-clock, 365 days a year, an appliance was more economical. And it was more responsive as well. John DeTore, Chief Investment Officer at Capitalogix explained that an appliance “allows us to get enough data to actually solve a problem in a statistically significant way and get enough processing done that we can get it done before the decision is behind us.”
The IBM Integrated Analytics System (IAS) offered benefits beyond performance. Its cloud functionality lets the team change deployment models as needed, and they can leverage offsite resources for functions such as disaster recovery. And its built-in data science tools such as Spark allow data scientists to build, train and deploy models all within the platform, using languages they know, without having to move data around.
A foundation based on data diversity and speed
Virtually any firm can say that data is their lifeblood. But for Capitalogix, the connection is direct and inescapable. Without data – all of it, all the time, of all types – they can’t find the edge that will help them find profit. As their CEO Howard Getson said, “there’s a difference between guessing and knowing and knowing is more profitable. If you know the answer, you can bet confidently and take decisive action, while other people are still guessing and therefore tentative. It’s a huge advantage in trading markets.”
It’s not just Capitalogix who sees the value in knowing the most profitable action to take in the moment. For an analyst’s perspective on appliances’ ability to provide quick insights, read IDC’s recent paper: “Delivering Hybrid Analytics at the Speed of Business.”
If you would like to discuss how your business can leverage the IBM Integrated Analytics System, schedule a no-cost, one-on-one consultation with an IBM data management professional today. z