Political intelligence website GovBrain uses an automated process to analyze political "events," then ties them to a ticker symbol as a way of determining how government decisions affect the financial markets.
The name, GovBrain, which founder Dr. Brent M. Eastwood admitted seemed to be "kind of an oxymoron, given political developments," is actually the application of technology, including machine learning, to a political intelligence industry that Eastwood says has been around for a long time.
GovBrain took second place in the 2015 Benzinga FinTech Awards. Eastwood spoke with Benzinga about the startup and its unique future in the growing world of financial technology.
Benzinga: What’s the history of the political intelligence industry upon which GovBrain is based?
Brent Eastwood: (It's) been around for a long time. You could probably say George Soros was the pioneer of political intelligence back in 1970. He traded based on information he knew from Europe – he knew some of the finance administers and politicians in Europe, so he started doing that.
That industry focused more on phone calls and conference calls and meetings and long reports, but wasn't able to grab the actionable trade in a systematic matter.
That's what we do. We systematize that whole "take the government information that's publicly available and legal" (and utilize it).
BZ: How does the process work?
BE: That (grabbing publicly available government information) is the first task the search engine does. Then it does another one based on political news sites like The Hill and Politico.
Usually these types of things impact a sector, so we do sector ETFs as best we can. Sometimes that's hard.
In addition, we do global macro. So, it's international, it's equities and bonds and currencies and commodities too.
BZ: Where does the prediction part come in?
BE: After a match is made with a ticker symbol or ETF, an automated sentiment analysis takes over. That's the prediction engine that we've trained with machine learning.
I train after every trading day or every week; I go back and confirm a prediction or correct it. The machine gets smarter over time.
BZ: How did you come up with this concept?
BE: I'm a former military officer. I have a PhD in political science. Having worked in state and local government, I got a good idea about how much data and information even medium-sized city governments and state governments put out.
I realized that as a political scientist it would be exciting to get into the financial markets, doing it in a way that combined my previous experience with political science and economic forecasting.
BZ: What makes GovBrain unique?
BE: I think the production of our own proprietary data, and the way we take huge government data, and we bring it in through a search engine.
Then it's parsed and goes through our sentiment analysis, our prediction engine. That leaves proprietary predictions.
Most people use big data because it's a byproduct of what they do. We use what we call little data – something I've been trying to coin.
BZ: So the predictions are the product?
BE: Our predictions are our main product. We aggregate them and do different things with them.
That's where the little data comes through. We probably do about 300 events a day that we predict.
BZ: Who makes up your user base?
BE: We wanted to reach a community that had no exposure to trading with government information, even though many folks on Wall Street have been doing it for a long time.
I wanted to educate retail traders and people who use (platforms like) StockTwits.
BZ: What’s the plan to monetize the platform?
BE: The business model is a licensing agreement for the data stream, our main product right now. That can go a couple of different ways. It could be per firm, it could be per seat.
We have quite a number of potential customers, as you could imagine, after the exposure we got from Benzinga Fintech Awards.
BZ: So Benzinga’s FinTech Awards were helpful then?
BE: Yes. Kyle and Jason did an amazing job, and it was so much fun just being there. We were just happy to be there.
What we had to do after the awards was kind of rate all our potential customers. It was like our top five list, and that became six then seven then eight. More people were coming in and we kept changing the list.
We finally found what we think is our number one customer, so we'll see how that goes. The demand from the (FinTech Awards) contest was definitely overwhelming.
BZ: What does the future look like for GovBrain?
BE: There are different things we can do. We can use mapping software, we can use Google Earth or Google Maps and show where we think if there's a crisis area where the contagion is, where it starts, like a patient zero thing.
Then being able to drag down your asset class either weekly or monthly. That's gonna be exciting. It's a new departure for me personally, and our team, because we're all "back end" people.
If it weren’t for our customer demand, we'd probably still be fiddling with the back end.
Related Link: Report: U.S. Government Considering Criminal Charges Against GM
BZ: Does this mean you are working more on the user interface?
BE: I've had people tell me, ‘Brent, you gotta stop messing with the machine learning. Let it go. Bring it to market.’
Our next step is defining that user experience so they can see what GovBrain has to offer, and make sure that we're not heading towards Armageddon, or we don't have any asset bubbles.
At the time of this writing, Jim Probasco had no position in any mentioned securities.
Image Credit: Public Domain
See more from Benzinga
© 2015 Benzinga.com. Benzinga does not provide investment advice. All rights reserved.