The Framework
Most stock research is a chat transcript. Ask a model about a company twice and you get two different
answers, neither of which you can compare against the last one. The framework exists to remove that:
every report here uses the same five sections, the same screens, and the same scoring.
That is the whole product. Not the length, not the charts — the fact that a report on FIGR can be laid
next to a report on VST and you can see precisely where they diverge.
Why a fixed framework
An equity research note is useful when you can tell three things apart:
- What the business is — factual, from filings.
- What you believe about the business — the qualitative judgment, where the real work is.
- What the price already assumes — the bridge between the two.
Most published research collapses these into one narrative, so a reader cannot audit the reasoning. This
framework keeps them apart at every step.
Step 1 — Business anatomy
Draw the skeleton before the leaves. Four questions, answered from filings rather than from the investor
deck:
- Company profile — what is actually sold, to whom, and how the business is structured.
- Business model and revenue mix — where revenue comes from, what margin each part carries, and where
the next one to two years of growth will come from. - Key financials — computed from SEC XBRL data by our own pipeline, with ratios derived rather than
purchased from a vendor. - Operating archetype — the classification that predicts everything downstream.
The three operating archetypes
DuPont forces the question: is this business high-margin and low-turnover, low-margin and high-turnover,
or leverage-driven?
| Archetype | Signature | What usually breaks it |
|---|---|---|
| High-margin, low-turnover | Gross margin above 50%, asset turnover low | Pricing power eroding; new entrants |
| Low-margin, high-turnover | Thin margins, revenue per asset high | Operating leverage in reverse; cost inflation |
| Leverage-driven | Returns produced by the balance sheet, not the income statement | Refinancing, rate sensitivity, covenant pressure |
The archetype is not a label. It determines which financial series is the leading indicator, and which
quarterly number would actually invalidate the thesis.
Step 2 — The investment case
Supply-demand structure
Every market sits in one of six structures, assessed over at least a five-year view:
| Structure | Read |
|---|---|
| Demand expanding, supply expanding | Early-stage. Many stories, high uncertainty, usually a high premium |
| Demand expanding, supply stable or shrinking | The golden structure. Demand durability is the question; supply contraction must come from an irreversible barrier |
| Demand stable, supply clearly shrinking | Mature oligopoly. Low headline growth, frequently mispriced |
| Demand stable, supply stable | Low-barrier industry. Look for differentiation or a new model |
| Demand stable, supply clearly growing | Seller's market turning buyer's. Look for new demand or the ability to squeeze a rival out |
| Demand shrinking, supply shrinking | The genuine end of an industry. Avoid |
Competitive intensity and the five moat types
Competitive intensity has three settings: low (pricing power, structure hard to disrupt), medium, and high
(price war). Low intensity comes either from an oligopoly equilibrium or from a fragmented market where
demand is highly differentiated.
The moat, if there is one, is one of five:
- Cost advantage — scale relative to rivals, regional economics, or a business model that structurally
lowers unit cost. - Non-market resources — exclusive assets, unique locations, relationships competitors cannot buy.
- Customer stickiness — brand, switching cost, network effects, and embeddedness in the customer's
workflow. The strongest and the hardest to build. - Administrative barriers — licences, patents. Powerful, but revocable; autonomy is low.
- A head start — operational efficiency or model innovation. Real, but the narrowest moat, because it
is the most copyable.
Pricing power has five levels: none (regulated), watching competitors, watching cost, watching the value
added, and watching supply scarcity. Levels four and five are where "small but strong" lives — and they
combine with wide replication room to describe the best businesses.
One distinction the framework insists on: a moat is protection; commercial value is the castle; investment
value is a third thing. A competitive advantage that never shows up as high return on equity or improved
business economics is a narrative, not an advantage.
Step 3 — Risks and financial red flags
The risk section is written to be specific rather than comprehensive. Each risk is stated with the
mechanism by which it would reach the numbers, and — where the evidence allows — the metric that would
surface it first.
The red-flag checklist is a separate, mechanical pass over the financials: receivables growing faster than
revenue, inventory building faster than sales, capitalised costs that would otherwise be expensed, related-party
revenue, share count drifting against reported earnings, and cash flow that persistently disagrees with net
income.
Step 4 — Preliminary value assessment
Five traits of a high-value business
| Trait | Test | US evidence |
|---|---|---|
| Large, monetisable value | Real social value that can be charged for, in an early-to-mid expansion phase | TAM, penetration, monetisation rate |
| High-quality business economics | Fits the cash-flow test: low capital intensity, high cash conversion, near-perpetual | FCF margin, capex/revenue, FCF yield |
| Value-expansion stage | Return on equity with room to rise, or high and durable, or a growing equity base | ROE height, durability, equity growth |
| High replacement cost and pricing power | Hard to replicate, pricing power at levels four or five | Gross margin, pricing behaviour, intangibles |
| Trustworthy, competent management | Governance, disclosure quality, incentive alignment | Filings, MD&A, call behaviour, incentive design |
R / N / g
| Element | Meaning | The test |
|---|---|---|
| R | Return on equity — level and how much room to rise | R below a future R′ implies value-expansion potential |
| N | How long the advantage can be sustained | Longer is better: five years, ten, or more |
| g | Rate of net-asset expansion | Asset-light businesses compound through retained earnings; asset-heavy ones through reinvestment |
R is supported by net margin, turnover and leverage. g is a property of the business archetype. N depends
on the long-run size of the market and how defensible the competitive advantage is. Together they locate the
company on its value-creation cycle.
Three valuation lenses
- Historical range — where this company's own multiples have sat across cycles.
- Peer comparison — what comparable businesses trade at, and why the discount or premium is deserved.
- Terminal value — what the business becomes if the growth rate persists long enough to matter.
Three scenarios
Bull, base, and bear, each with its own set of stated assumptions — never a single number pretending to be
a forecast.
Step 5 — Valuation range under stated assumptions
The range is built three ways and then converged:
- EV/EBITDA peer scenario — applying a multiple range to forward EBITDA.
- P/E comparison — applying a multiple range to forward earnings.
- P/S comparison — applying a multiple range to forward revenue, with the margin path stated separately.
Where the three overlap, that is the range. Where they diverge, the divergence is itself information: it
usually means one of the methods is anchored to something the others are not, and that gets written down
rather than averaged away.
What this framework deliberately does not do
- No price target. Every conclusion is a range under stated assumptions, because the assumptions are
the part worth interrogating. - No position sizing, and no entry instructions. The output is a research verdict and a monitoring
framework, not an order. - No "buy", "sell", or "hold" rating. A verdict says whether a business deserves your time — worth
researching, worth monitoring, or pass for now — and what would change it.
That is a deliberate constraint, not an oversight. It also happens to be the thing that makes the reports
usable: a reader can act on the analysis without inheriting someone else's risk tolerance.
How the data is sourced
| Input | Source | Method |
|---|---|---|
| Financial statements | SEC EDGAR XBRL | Automated pull; ratios recomputed from raw tags, not purchased |
| Segment and revenue mix | 10-K / 10-Q filings | Read from the filing |
| Market and industry data | EIA, industry associations, public research | Cited per report in the appendix |
| Peer multiples | Public market data | Sourced and dated per report |
Two dates are printed on every report and never blurred: the financial period the analysis is based on
(filing type and date), and the valuation date the multiples are struck at. They are different things,
and a report that hides the difference is telling you something.
The verdict
Every report closes with one of three verdicts, plus the conditions that would change it:
| Verdict | Meaning |
|---|---|
| Worth researching | The business and the price both stand up. Worth spending an afternoon on. |
| Worth monitoring | The business is real, but the current price already discounts most of it. Waiting for a better price or clearer evidence. |
| Pass for now | The business or the price structure does not work right now. Not worth the time. |
A verdict is a statement about attention, not about transactions. It is the filter the whole product is
built around: decide first whether a company is worth deep work, then decide whether the work is worth
your money.