Chapter 1

What FactSet is, and why the price has halved

FactSet sells financial data, analytics and workflow software to the investment industry on multi-year subscriptions, and it has grown revenue for 45 consecutive years and diluted EPS for 29. Yet the shares have fallen from $495.72 in November 2024 to $262.46, roughly halving, as organic growth slowed to mid-single digits and the market began pricing generative AI as a threat to the terminal-and-data model. This chapter orients a cold reader and fixes the question the report exists to answer.

Sources: price, market cap and drawdown computed from daily market prices (all-time-high close $495.72 on 14 Nov 2024; $262.46 on 16 Jul 2026); revenue and EPS from the FY2025 Form 10-K [1] [2]; ASV from the FY2025 10-K [3].

The business, in plain terms

FactSet (Norwalk, Connecticut; founded 1978; fiscal year ends 31 August; NYSE: FDS) aggregates market, fundamental, and alternative data and delivers it through a workstation, feeds, and analytics that portfolio managers, sell-side bankers, wealth advisers, and corporate teams use to research, model, and monitor investments. Revenue is almost entirely recurring: clients contract for a defined set of data and users, and the sum of those annualized contracts is the company's headline operating metric, Annual Subscription Value (ASV), which stood at $2,405.6 million at the end of FY2025 [4].

The switching costs show up in the numbers a value investor cares about first. At FY2025 year-end FactSet served 8,996 clients and 237,324 users [5], retained more than 95% of ASV over the trailing year, and kept 91% of clients by count [6]. Its customer base is the industry's core: 95 of the top 100 global asset managers, 43 of the top 50 investment banks, and 37 of the top 40 wealth firms, with nine of its ten largest clients on the platform for more than 20 years [7].

That embeddedness is what has powered a rare compounding record: 45 consecutive years of revenue growth, 29 consecutive years of EPS growth, and 27 consecutive years of dividend increases [8]. Users have more than doubled since FY2018.

Source: user counts by fiscal year from FactSet segment KPI disclosures, consolidated into the FY2025 10-K figure of 237,324 [9].

The economics: high-margin, cash-generative, lightly levered

FactSet converts that recurring revenue into wide margins. FY2025 revenue was $2,321.7 million and operating income $748.3 million — a 32.2% operating margin, up from 25.8% in FY2022 [10] [11]. Diluted EPS rose 11.8% to $15.55, and the company returned $460.4 million to shareholders through buybacks and dividends during the year [12].

Source: consolidated statements of income, FactSet FY2023 10-K (FY2021–FY2023) [13] and FY2025 10-K (FY2024–FY2025) [14].

The balance sheet is conservative for a business that carries acquisition debt: about $1.37 billion of total debt against roughly $0.34 billion of cash and $726 million of operating cash flow in FY2025. Interest is covered many times over, and net leverage sits near one-and-a-half times cash earnings. For a reader who wants the probability of bankruptcy near zero, FactSet clears that bar comfortably — the debt was taken on to buy the CUSIP Global Services franchise, not to plug an operating hole, and it is amortizing.

Why the market marked it down

Two things changed at once. First, growth decelerated. Reported revenue growth fell from 15.9% in FY2022 to 5.4% in FY2025, and — stripping out the CUSIP acquisition that flattered FY2022–FY2023 — organic ASV growth was 4.8% in FY2024 and 5.7% in FY2025 [15] [16]. A franchise the market had valued as a mid-teens compounder was resetting to mid-single digits.

Source: year-over-year growth computed from consolidated revenue in the FY2023 and FY2025 10-Ks [17] [18].

Second, the story got harder to underwrite. Generative AI cuts both ways for a data vendor — it can lower the cost of assembling data and answering research questions, which threatens the premium a workstation commands, even as FactSet embeds the same tools in its own products. And in June 2025 the board replaced Phil Snow, a 30-year insider and CEO of a decade, with Sanoke Viswanathan, a 15-year JPMorgan executive and the first outsider to run the company in its modern history [19]. A first-time outside CEO arriving as growth slows is a governance change a buyer has to price.

The result is a sharp de-rating. At its November 2024 peak the stock traded near 35.6 times trailing earnings; at $262.46 it trades at 16.9 times FY2025 EPS and about 14.7 times the $17.81 the analyst consensus expects for FY2026 — a market multiple for a business that spent years commanding a large premium.

Source: calendar year-end closing prices from daily market data; the intraday-basis all-time-high close was $495.72 on 14 November 2024.

The through-line this report follows

FactSet is a fallen star: a best-in-class financial-data franchise, with 95%-plus ASV retention and a 45-year growth record, whose shares have roughly halved from their late-2024 peak as organic growth settled into the mid-single digits and the market began pricing generative AI as a structural threat to the data-and-terminal model. The question this report exists to answer is whether the de-rating has handed a value investor a genuine margin of safety — in a business whose bankruptcy risk is near zero — or whether roughly 15 times forward earnings still overpays for a franchise that no longer compounds the way its record implies.

Two facts frame that question honestly. In FactSet's favor: the moat is visible in the retention and client-tenure data, the cash conversion is high, and the balance sheet removes the tail risk this reader most fears. Against it: FactSet is not founder-run — insiders own little, and a first outside CEO is arriving precisely as growth slows — so the "skin in the game" this reader prizes is thin, and the AI question is genuinely unresolved. What would settle it is evidence over the next several quarters on whether organic ASV growth stabilizes or keeps sliding, and whether AI shows up as a tailwind in FactSet's own pricing or as a discount in its clients' willingness to pay. The chapters that follow test each leg in turn.