Methodology

Intrinsic value is the present value of unlevered free cash flow to the firm, discounted at the weighted average cost of capital. FCFF is net operating profit after tax, plus depreciation and amortisation, less capital expenditure and the change in net working capital.

A net income or free cash flow to equity model omits the cash a company must spend on capital expenditure and working capital. Those are real uses of cash, so leaving them out overstates what is available to shareholders.

Stock-based compensation is deducted as a real economic cost. It is non-cash under accounting rules, but it transfers value to employees and dilutes holders, so treating it as free would overstate intrinsic value.

WACC is built from the capital structure and the cost of each source. The cost of equity comes from CAPM using a live risk-free rate, the company's own beta, and an equity risk premium. The cost of debt uses the company's interest expense against its reported borrowings, tax-effected at the effective rate.

Beta, the risk-free rate and the equity risk premium are analyst-calibrated inputs rather than fetched values. They are surfaced in the WACC breakdown panel in the workbench so you can see exactly what went into the rate.

The discounted cash flows give an enterprise value. Equity value is that figure adjusted for net debt or net cash, minority interest, and any preferred stock, then divided by the share count resolved from the listing structure. Net cash treats marketable securities as cash, and the per-share figure divides by diluted shares outstanding.

The bridge is shown in full in the DCF schedule, so a difference between the enterprise value and the equity value can be traced line by line rather than taken on trust.

Every model runs three scenarios: base, bull and bear. They differ in the growth fade, margin path, capital intensity and working capital assumptions. The implied price for each is shown against the current market price, which is what the header strip reports.

The base case is generated, not polled. Each driver starts from the company’s most recent run rate and tapers across the forecast years, so Base can carry a high-velocity year-one through. It is the central path we would assign the filings, not a consensus of analyst estimates; bull and bear bracket that path rather than extending the method itself.

Driver overrides you set in the panel are recorded against the model-generated baseline, so any change can be reverted and the original value is never lost.

Financial statements are read from public filings and market data providers:

• United States: SEC EDGAR company filings, primarily Form 10-K and 10-Q.
• India: published financial statements for NSE and BSE listed companies.
• Share counts and market prices: daily closing quotes from free market data providers, refreshed once per day.

Reported line items are mapped to a canonical taxonomy before anything is calculated, so a line labelled differently between two filings still lands in the same place in the model.

Prices are a daily close, not a live tick. The KPI bar labels every market price with the date it was taken and the source it came from, and a provider that fails leaves the last known good quote in place marked stale rather than substituting a placeholder.

Because a close is up to a day old, treat the implied-versus-market difference as a comparison against the last close, not against a real-time quote.

Every model runs a set of structural checks before it is served: that the forecast balance sheet balances, that the cash flow statement articulates with the income statement, that peer multiples are measured on the same basis as the company they are compared to, that the enterprise bridge uses the net figure from the right balance sheet date, and that the WACC in the Excel export reproduces the number in the API.

The results are in the QA panel in the header. A model that fails a check on its inputs is not presented as a valuation at all: no headline price, no implied-versus-market comparison, and nothing in the share card or the meta description. 10 of the 14 pre-built models publish. The other 4 open with their full statements, their bridge, their WACC build and the failing checks named, and with no price.

That is the opposite of listing the failures under a headline and asking the reader to discount one number, which is what this page used to say the product did. A reader cannot see which figure to discount, and the headline is the only thing most of them read. Withholding the evidence as well would be worse, because then the refusal could not be checked, so the audit stays and the conclusion does not.

Two rules about the checks themselves. A check that cannot run is marked skipped rather than quietly counted as a pass, and a check that examined nothing fails rather than passes, because an unverified check that reports success is believed, and that is the one outcome worse than a check that did not run.

Stated plainly, because a model that overstates its reach is not useful:

• Not every listed company has a model. Where financial statements cannot be sourced from the filings read, no model is built and the ticker page says so rather than showing an empty result.
• Financial sector companies are out of scope. Bank and insurance balance sheets do not fit an enterprise-value bridge, so they are excluded from the universe entirely.
• Forecasts are the model's own arithmetic, not consensus estimates. No sell-side figure is used anywhere.
• Calibrated inputs such as beta and the equity risk premium are analyst judgement, not measurement. They are shown so they can be disagreed with.
• Saved models are stored in your browser only. They are not uploaded, not synchronised, and clearing site data removes them.

Valence computes a valuation from stated assumptions and public filings. The output is the model's own arithmetic, not a view on whether to buy or sell anything. Assumptions are yours to change, and the model shows what each change does. Check the source filings before relying on any figure here.

Nothing on this site is investment advice. Past performance and modelled forecasts do not indicate future results.

Spotted a figure that looks wrong, or want a ticker covered? Send a note.