Trace how growth, margins, reinvestment and the cost of capital change the value of a business.
What must be true about a company's economics for a valuation to make sense? A share-price estimate alone conceals that question. IntrinsicLab follows the assumptions through annual cash flows, financing costs and a transparent enterprise-to-equity bridge, then shows how readily the answer changes.
Revenue growth can create value or consume it. A company that must invest heavily to grow may generate much less cash than its income statement suggests. I wanted to make the connections between profitability, reinvestment, return on capital and valuation inspectable in a working product.
FCFF = EBIT − cash taxes + D&A − CapEx − change in operating working capital.
These cash flows belong to all capital providers, so they are discounted at WACC. CAPM estimates the cost of equity; market-value financing weights blend equity and after-tax debt costs. Terminal value can use a Gordon-growth perpetuity or an EV/EBITDA exit multiple.
In the Gordon model, stable reinvestment is NOPAT × growth / ROIC. Growth is not free. Enterprise value is reconciled to equity after debt, preferred capital, noncontrolling interests and nonoperating cash. Money and shares are both in millions, so their ratio is ordinary currency per share.
- Editable annual growth, operating margins, D&A, CapEx and working-capital drivers.
- CAPM and market-value WACC with an explicit tax-shield assumption.
- DCF using stable reinvestment or final-year EBITDA exit multiples.
- WACC × terminal-growth / exit-multiple sensitivity with invalid cells explained.
- P/E, EV/EBITDA, EV/sales and P/book for imported peers; undefined multiples remain unavailable.
- JSON model import, annual CSV normalization, provenance, scenario comparison and reproducible evidence export.
The browser derives every result from the Python engine. There is no invented quote feed or AI-generated financial number. The included fictional company is visibly labeled DEMO DATA. Imported source information is user-asserted and retained, not certified as audited.
The included Northstar Industrial teaching model calculates USD43.26 per share at 9% WACC and 2.5% terminal growth. About 76% of enterprise value comes from the terminal period. Across the displayed WACC/growth combinations the range is USD30.95–75.40. These are observed outputs of the versioned fictional assumptions, not market data or a confidence interval.
The interesting result is the dependence on long-term economics: a plausible-looking central number can coexist with substantial assumption risk. Recalculate the model at a higher WACC, or lower terminal ROIC, and investigate where the value disappears.
Python3.12+ and Node22+ for browser checks. No API key, database or Docker daemon is needed.
python3 -m venv .venv
.venv/bin/python -m pip install -r requirements.lock
.venv/bin/python -m pip install --no-deps -e .
.venv/bin/python -m uvicorn intrinsiclab.api:app --host 127.0.0.1 --port 8016Open the workbench or OpenAPI docs. Use Import model for a JSON assumption document or an exported evidence file. Evidence import always recalculates the assumptions. Evidence & methods accepts annual CSV. The comparables screen downloads an exact input example.
Typed inputs → annual normalization → pure financial functions → analysis service → FastAPI → browser workbench.
See architecture for boundaries and data dictionary for units. Browser case snapshots stay in local browser storage. The API is stateless and intended for personal local research.
npm ci
npx playwright install chromium
scripts/verify.sh
.venv/bin/python scripts/check_secrets.pyValidation covers hand-calculated FCFF/DCF identities, reinvestment, loss taxes, negative equity, denominator failures, chronological annual imports, API errors and request limits. Browser tests exercise changing WACC, clearing invalid results, saving/restoring cases and mobile layout. CI runs tests, lint, Python type checking, browser checks, package build and a tracked-file secret-pattern scan. Python transitive versions and Node dependencies are locked.
Annual nominal forecasts use end-of-year discounting, one currency and constant tax assumptions. No loss carryforward is modeled. WACC assumes usable tax shields; the analyst must supply appropriate market financing values. Terminal margins and ROIC require a credible transition from the explicit period. Banks, distress and real options need richer models. No live issuer information or personalized recommendation is supplied.
Read methodology and limitations before interpreting results. A sensitivity interval is a set of conditional scenarios, not a probability statement.
How should operating margins and incremental ROIC converge together? When does an exit multiple imply an implausible reinvestment policy? How much do stub periods, leases, stock compensation and tax-loss utilization change the equity bridge?
MIT licensed. Contributions should preserve the distinction between evidence, assumptions and calculated interpretation.