The LSE Directory / Alternative data
M Science is a New York based alternative data research firm that sells data driven investment research to institutional investors. Its core product combines analyst curated written research, published two to three times per quarter on individual companies, with interactive dashboards that visualize company and industry level KPIs; the firm tracks more than 1,400 unique KPI metrics forecast quarterly across roughly 1,400 public and private companies, delivered through downloads, API, S3 and Snowflake Share, plus more than 600 programmatic data endpoints. In 2026 it added an AI layer to this stack: an "Insight Copilot" called Maddie, launched in February 2026, and a Unified Data Model paired with an MCP Server, launched in June 2026, that exposes its research and datafeeds directly inside client AI tools such as ChatGPT and Claude. The firm specialises in triangulating multiple alternative data types, originally anonymised consumer transaction records and now also digital purchase, web traffic and technographic data, into sector research rather than selling raw data alone; this analyst plus dashboard plus feed combination is aimed at hedge funds, asset managers and corporates that want both a narrative view and machine readable KPIs, distinguishing it from data vendors that sell only raw feeds or only written research. Coverage spans consumer retail, industrials and machinery, financials, TMT, semiconductors, software and airlines, across the US, UK, Europe, APAC, Australia and Latin America, with offices in New York and Hong Kong. The firm traces its lineage to Majestic Research, founded in 2002 as one of the first research houses built on large scale commercial transaction data rather than traditional sell side analysis; it was acquired by ITG in 2010, then sold on to Leucadia National in 2016 and rebranded M Science, adopting artwork and typography that echoed the Majestic Research logo. It has operated as a subsidiary of Jefferies Financial Group (NYSE: JEF) since the Leucadia deal. In January 2023, M Science sold its TMT research and data assets to Guidepoint, a transaction the firm described at the time as sharpening its focus on consumer and industrial sector data; its current site nonetheless lists TMT among its active research verticals alongside consumer, industrials, financials, APAC and airlines. The firm has repeatedly won "Best Alternative Data Provider" at the HFM US Technology Awards, including in 2022, its fourth consecutive win and eighth award in four years at that point. Leadership includes CEO Michael Marrale and COO Joseph Napoli. Most recently, M Science has pushed further into AI native distribution of its research: the June 2026 Unified Data Model standardises what it describes as more than 1,440 KPIs across roughly 1,400 companies with near real time updates, and its MCP Server gives clients tool based, programmatic access to that data from inside large language model applications, positioning the firm's data as infrastructure for automated investment workflows rather than only a research product consumed by human analysts.
