This Startup Just Got $10M to Predict Politics with Tech
What can a company do to ensure their business plan is not bushwhacked by a new tax or regulatory scheme? For years, the answer has been to hire lobbyists and law firms that keep tabs on what politicians are up to in Washington and state legislatures. But now, a 23-year-old whiz kid whose startup raised a Series C funding round is making waves with a new approach.
The startup is called FiscalNote and instead of relying on lobbyists, it uses technology to scope for new laws and regulations across the country. And when it discovers something, FiscalNote uses big data and machine learning techniques to predict if a proposed bill will affect an industry.
In practice, this means FiscalNote deploys thousands of Python scrapers to suck up data from government websites, and then relaying important information to its clients. For instance, the tool may spot a new FDA proposal that could affect the pharma industry or a state bill in Kentucky that could restrict the telecom sector.
According to CEO Tim Hwang, his company’s data-based approach can predict the fate of a bill with 94% accuracy, and is changing the way companies approach politics.
Sign up for Data Sheet, Fortune’s technology newsletter.
“Think of Washington and the influence peddling and information peddling that’s gone on for 300 years. For folks who are aggregating info, and understand affect law on business—computers do that very well,” Hwang told Fortune in a recent interview.
He should know. At the age of 16, Hwang used his data talents to help President Obama win the Iowa caucuses, and then to get elected himself to a county government position in Maryland. In 2013, he launched FiscalNote and has since raised money from big-name investors like Dallas Mavericks owner Mark Cuban and Yahoo founder Jerry Yang.
Today, FiscalNote is proving especially popular with the health care sector and, fittingly, the tech industry where companies like Uber, Lyft, and Coinbase (a bitcoin startup) use it.
On Tuesday, the company announced a $10 million Series C funding round, led by Visionnaire Ventures and Green Visor Capital, which Hwang says it will use to expand Asia and European operations, and to invest in new product lines in law and compliance. He added the company will safely reach profitability in the next few quarters, and that is revenue is in the “upper seven digits” and is growing year-over-year at 900%.
The Science of Predicting Politicians
FiscalNote’s data-based approach to law making is an important tool for companies, but also highlights features of the political process itself. One of the most interesting—or perhaps discouraging—of these is the hundreds of thousands of bills proposed every year that are basically meaningless.
Hwang gave the example of the dozens of bills introduced by Republicans in Congress to repeal the Affordable Care Act. Think what you want of the President’s health law but it’s hard not to agree with FiscalNote’s data tools, which will examine a new bill and conclude (in Hwang’s words), “If you’ve voted on this 47 times, it won’t pass the 48th time.”
In this sense, the startup’s software is useful because it can save companies from worrying about all the proposed laws that are mere political laws, and focus on the bills that could actually pass. In the latter case, FiscalNote looks for signals such as the presence of key co-sponsors to predict a law’s likelihood of success.
“The amazing thing about politicians is they vote their preferences year after year. You can anticipate how polticians will react to issues they’ve seen many times before, and combine with social network analysis and raw computational capacity.”
But how can FiscalNote be so sure its technology, which uses a MongoDB database paired with a Hadoop machine-learning cluster, doesn’t get it wrong?
Hwang acknowledged that FiscalNote is hardly fool-proof, but says that its 94% accuracy rate is far better than conventional political intelligence tools.
For more on startups, watch:
“There are tons of black swan events in politics. But now, the baseline is journalists who are often lucky or they’re wrong. In a world of computational power, it doesn’t make sense to make gut decision about something as important as laws and regulations.”
In this sense, Hwang regards the use of analytic tools to predict politics as a logical outgrowth of so-called “legal tech,” in which a swath of firms sell predictive tools to help companies assess the outcome of court cases.
But some challenges remain, including obtaining data from local and state legislatures, whose IT platforms can make the federal government look positively modern in comparison. Meanwhile, FiscalNote also has to contend with industry giants like Bloomberg and Thomson Reuters, which offer political intelligence and compliance tools of their own.
Hwang, though, is confident that FiscalNote’s so-called ingestion engine is a “secret sauce” that will allow the startup to eat more marketshare.
“We have the best tech in the market, period. Our IT is faster and higher quality. I worry less about Lexis and Thomson and more about early stage competition.”
Correction: an earlier version of this article stated Hwang was elected to a state government position; it was a county position.