Berkshire requires a 25–50% discount to intrinsic value before buying.
Buffett Quality Checklist
✓ROE >15% consistently (≥7 of last 10 years)
✓Free cash flow positive (≥8 of last 10 years)
✓Conservative leverage — Debt/Equity below 1
✓Revenue growing at CAGR >5%
✓EPS growing at CAGR >5%
10-Year Financial History — SEC EDGAR 10-K Filings
Year▲
Revenue▲
Net Income▲
FCF▲
Owner Earnings▲
ROE▲
Net Margin▲
LT Debt▲
Cash▲
2016
$1.1B
$338.8M
$283.4M
$329.1M
65.5%
30.1%
$300.0M
$228.4M
2017
$1.2B
$258.3M
$283.7M
$269.7M
46.1%
21.1%
$575.0M
$194.7M
2018
$1.4B
$267.1M
$352.1M
$290.9M
50.8%
19.8%
$574.8M
$208.6M
2019
$1.4B
$352.8M
$367.8M
$353.9M
52.5%
24.6%
$574.2M
$359.8M
2020
$1.5B
$372.9M
$428.2M
$352.9M
41.6%
25.0%
$574.4M
$585.6M
2021
$1.6B
$399.6M
$493.9M
$402.7M
39.3%
25.1%
$574.5M
$681.9M
2022
$1.8B
$396.9M
$487.1M
$432.4M
29.8%
21.5%
$2.0B
$503.3M
2023
$2.1B
$468.2M
$584.8M
$512.8M
28.9%
22.4%
$1.6B
$425.4M
2024
$2.2B
$537.1M
—
—
28.1%
24.4%
$1.4B
$423.0M
2025
$2.3B
$597.0M
—
—
27.3%
25.7%
$1.4B
$337.7M
Warren & Charlie
Buffett / Munger — quality, moat & valuation
FACTSET RESEARCH SYSTEMS INC (FDS) — Investment Memo
🐂 The Bull Case (Warren's voice)
The "Toll Bridge" Reality: This isn't just software; it’s an integrated utility. Clients don't buy FactSet because they want to; they buy it because their entire day—from Excel modeling to portfolio attribution—is hardcoded into the platform.
Pricing Power: When you provide 90% of a client’s daily workflow, you aren't an expense; you’re an infrastructure cost. They have successfully grown revenue from $1.1B to $2.3B over the last decade, proving they can raise prices without mass churn.
Capital Light Compounding: They don't need to build factories or buy inventory to grow. With an ROE of 27.3%, the business generates excess cash that management can use for buybacks or smart, tuck-in acquisitions, creating a feedback loop of compounding value.
The Target: We pay for certainty. We like businesses that don't need to "reinvent" themselves every five years. FactSet is the rare company that is just as relevant today as it was in 2016, simply by doing the same thing slightly better each year.
🐻 The Bear Case (Charlie inverts)
The Commoditization of Data: The greatest threat is not a competitor, but the structural decline of "proprietary" data value. If LLMs and AI can pull, normalize, and interpret public financial data as well as a human-curated workstation, the "toll bridge" loses its gate.
Headcount Sensitivity: FDS is a proxy for the size of the professional financial workforce. If the industry shifts entirely to passive, algorithmic, or AI-driven management, the seat-based revenue model dies. We do not want to own a business that is structurally tied to the labor force of Wall Street.
The "Legacy" Trap: Every piece of proprietary software eventually becomes legacy code. If a new, agile, AI-native competitor builds a workflow that makes the FactSet workstation feel like a clunky spreadsheet plugin, the switching costs will vanish overnight.
💰 Valuation & Margin of Safety
Our DCF model indicates an intrinsic value of $372.81 per share. Because we demand a margin of safety for the risks outlined above, we are unwilling to pay the current market premium.
Intrinsic value estimate: $372.81 per share
25% margin of safety entry: $279.61
50% margin of safety entry: $186.41
Current Market Assessment: The stock is currently trading at a significant premium to our intrinsic value estimate, placing it firmly in the "Expensive" category. We are unwilling to pay for growth that has already been priced in by the market.
Verdict: [PASS]
FactSet is a wonderful business, but the current price leaves us with no margin of safety to protect against the inevitable technological shifts in financial services. We insist on buying at a discount to intrinsic value, and at these levels, we are paying a premium for a future that is becoming less certain. We will keep this on our Watch list, but for now, we leave it to others.
Other Analyst Views· Lynch · Damodaran
Research Notes· Money Model · Moat · Financials · Management
Data sourced from SEC EDGAR XBRL filings (10-K only). For educational purposes — not investment advice.