Thinknum is a startup with the mission: disrupting financial analysis.
In his work as a quantitative strategist at Goldman Sachs, Thinknum co-founder Gregory Ugwi saw firsthand the trials and tribulations financial analysts went through to digest companies’ financial reports and then build their own research reports about their expectations for future performance based on past numbers. The U.S. SEC’s mandate that companies disclose their financial data using XBRL (eXtensible Business Reporting Language) was supposed to help them, as well as investors of all stripes and sizes that want to better understand what's going on at the companies they're interested in.
“The SEC has mandated that all companies have to release their numbers in a machine-readable format, and that’s XBRL (eXtensible Business Reporting Language),” says Ugwi. The positive side of that is that anyone can now get the stats on companies from Google to Wal-Mart, but the downside is that by and large, they can’t do it in a user-friendly way.
“It’s just terabytes of XML documents dumped into an FTP site,” says Ugwi. And financial researchers, he says, by and large don’t want to know what XML is, never mind worry about having to have software built to talk to use the SEC’s database of corporate filings and pull the XBRL-tagged data they need. So analysts and other investors often turn to third parties to aggregate and model it for them, assuming they can afford that option.
Ugwi thinks there's a better way. “We see a huge opportunity to build an open web platform – XBRL data at the back-end but at the front-end users can collaborate with each other to build the best financial models,” he says. “You don’t need a centralized system of build and research models. Now you can have a distributed framework for building financial models for companies.”
Gain Insight, Leverage Collaboration
With Thinknum’s financial analysis engine, semantic technology transparently connects over 2,000 data sources for information ranging from company-specific fundamental data to market data that can be analyzed with the help of machine learning, quant trading and other algorithms related to statistical functions such as curve trades and regressions. Where semantic technology is powerful in the form of dealing with XBRL is that users can get this data directly, bypassing intermediaries that interpret and supply financial data, he says.
For a company’s total revenue, for example, users can enter a ticker symbol and Thinknum will know to plot that over time, taking advantage of the machine-readable XBRL format in obtaining that data from the SEC database.
While anyone can financial build models on their own for analyzing cash-flows or other financial data, the value-add of Thinknum is both in making this data user-facing and enabling users to piggyback off each other’s work leveraging machine-readable data, Ugwi explains. “Once you have open and accessible data, the interesting thing is users can collaborate with each other. Financial models depend on each other," he says. “To value Apple, for example, you must be able to value Apple’s options, so you need to hook into someone else’s model that values options correctly.”
The stickiness of the platform lies in creating that network effect of model collaboration with others using Thinknum’s platform. In addition to openly and freely sharing these models, Ugwi says the platform also can accommodate the big investor banks that want to pay to use the platform for their teams’ private models. “But we really are focused on the open platform and growth there,” he says.
He thinks there will be growing interest in sharing models freely by those researchers who want to build reputations for their expertise in doing so. “People will always want to share to become more credible – there’s a lot of value in others’ using your model.”
Thinknum recently got on board with 500 Startups, which provides early-stage companies with up to $250,000 in funding, a startup accelerator program, access to startup mentors and more.