Credit card transactions, smartphone app usage, location data, satellite imagery, and scraped web pages now feed the investment decisions of hedge funds and private equity firms. Vast data sets and the computing power to analyze them can offer an information edge, but some of the most predictive data may also be material nonpublic information (MNPI), raising questions under the federal securities laws that regulators and fund compliance teams are still working through.
In this TalksOnLaw interview, Joel Cohen speaks with Igor Rozenblit, the former long-time co-head of the private funds unit of the U.S. Securities and Exchange Commission (SEC), about what alternative data is, when using it can cross the line, and how investment managers should manage the risk.
What Counts as Alternative Data
Rozenblit explains that alternative data is essentially any data obtained from non-traditional sources, as opposed to regulatory filings such as 10-Ks and 10-Qs or services such as Bloomberg. Its use began with fundamental long-short equity hedge funds and has spread across private equity and the broader alternative investment industry. There is no government definition; to the SEC, he says, data is just data, and what matters is whether it is MNPI.
When Is Alt Data MNPI?
Whether a data set is MNPI is a hazy question with no bright-line rules, but Rozenblit offers rules of thumb. Data that anyone could obtain, such as imagery from satellites orbiting the Earth, is less likely to be MNPI, while data that is not publicly accessible, such as individual credit card transactions or app usage data, carries more risk. His advice: unless you are comfortable defending the position that a data set is not MNPI, treat it as if it is.
The App Annie Case
Rozenblit walks through the SEC's case against App Annie, an app analytics company that told hedge fund clients its estimates were derived from aggregated, anonymized data, when in fact it used confidential app performance data obtained from the companies themselves. The SEC charged App Annie under Section 10(b), the provision typically used in insider trading cases, based on those misrepresentations. He explains why the SEC pursued the data vendor rather than the funds that bought the data, and the dilemma facing a fund that receives data that looks too good to be true.
Section 204A and Vendor Due Diligence
For fund managers, Rozenblit says, the greater risk is not insider trading, which requires proof of intent, but Section 204A of the Investment Advisers Act, which requires advisers to maintain policies and procedures reasonably designed to prevent misuse of MNPI and which, in his experience, is the provision SEC examiners cite when alternative data issues arise. He recommends identifying high-risk data vendors and confirming that each has its underlying sources' permission to sell the data for trading, either through a representation and warranty or by reviewing the vendor's contracts with its sources.
Web Scraping, CAPTCHAs, and Logins
As funds shift from buying vendor data to building in-house data science teams, web scraping is growing. Rozenblit explains why scraping public websites is generally lower risk but why clicking "I am not a robot" or scraping behind a login in violation of a site's terms of service could be characterized as deception under Section 10(b) or Section 17(a) of the Securities Act. He describes industry practice at the time as scraping past CAPTCHAs but not past logins, and why internally generated data, which comes without a vendor's representations or compliance program, can carry more risk than purchased data.
Igor Rozenblit is the founder and managing partner of Iron Road Partners, a regulatory compliance consulting firm. He served as the co-head of the private funds unit of the SEC from 2010 to 2021.
What to Know Now
App Annie and its founder, Bertrand Schmitt, settled the SEC's charges in September 2021, agreeing to pay penalties of $10 million and $300,000; the company rebranded as data.ai and was acquired by rival Sensor Tower in 2024. In an April 2022 risk alert, the SEC's Division of Examinations identified alternative data as an area where advisers' MNPI policies often fell short, citing inconsistent diligence on data providers and a lack of procedures for acting on red flags, while noting that alternative data does not necessarily contain MNPI. The SEC has continued to bring Section 204A cases against advisers whose MNPI policies were inadequate or not enforced, even without alleging any illegal trading. Questions about scraping have largely been addressed in contract and computer-access litigation rather than securities cases: in hiQ Labs v. LinkedIn, the Ninth Circuit held in 2022 that scraping publicly available data likely does not violate the Computer Fraud and Abuse Act, but the district court later found that hiQ had breached LinkedIn's user agreement, including by using fake accounts to scrape behind logins.
