Signal Ledger LabSignal Ledger · Our methodology
We value each business at FY2030, then discount that value and the cash generated along the way back to today.
One directional scenario: management delivers its published targets, with research estimates where targets are missing. We do not combine bear, base and bull cases or assign probabilities to them. This is a way to test what must be true, not a promise of what will happen.
Start with the source
Each company gets one Signal Ledger valuation. A company promise, a research inference and our own judgment should still be easy to tell apart inside it.
We start with public long-term targets, the latest guidance and reported results. A target needs a public company source, such as a filing, earnings call or investor day.
Revenue follows those targets; fixed rules fill the gaps. Analyst forecasts and peer valuation multiples do not set this case's value.
When management gives no usable long-term target, we can use cited consensus forecasts, explicit operating assumptions and the economics of comparable companies at a similar scale.
We explain the inference and its sources. Cyclical businesses use a mid-cycle view rather than assuming peak conditions last. These are research assumptions, not management promises.
Where the evidence leaves room for judgment, such as a business that has no revenue yet or a discount rate for a company with little trading history, we make the call ourselves.
Each call is part of the one published valuation. The company report lists every call with its reason and date; the same engine and validation rules apply.
Follow the economics
Revenue growth is only valuable if the business can turn it into cash after the investment it needs.
For software, consumer, hardware and similar businesses, revenue × operating margin gives operating profit. We deduct taxes and capital spending and add back depreciation to estimate cash flow.
AI and cloud infrastructure must buy equipment before it earns revenue. Beyond committed near-term spending, capital expenditure follows next year's demand, the equipment already in service and its useful life.
We check whether the equipment earns more than it costs. Faster growth can consume cash and destroy value if the unit economics do not work.
A common set of rules
A dollar received later is worth less than a dollar today. The discount rate captures that required return.
Cost of equity = normalized Treasury yield + β × 5%
The risk-free input is the 5-year average of the 10-year U.S. Treasury yield, rather than a single day's rate. Measured beta is adjusted toward the market average and initially limited to 0.8–1.2. For operating companies with debt/equity ≥ 0.5 or net debt/EBITDA ≥ 2.5, we relever the 0.8 floor using after-tax debt/equity, keeping the higher beta up to 1.6. Financial-company equity models are exempt.
Illustration, not a live rate: with a Treasury average of 3.85% and beta at its 1.2 cap, the rounded cost of equity is 9.85%. Leverage can raise beta above this ordinary cap; the Treasury average also changes over time.
Operating cash flows use the weighted average cost of capital (WACC), blending equity and after-tax debt costs. When an infrastructure line has negative initial cash flow and company net debt rises, the build-out premium increases proportionally from zero to 1 percentage point as the initial net debt increase reaches 2% of equity market value. The premium is capped there. Cost of equity is rounded to 0.01% steps; WACC is not rounded. The build-out premium can put WACC above the illustrated equity cost.
We use near-term guidance and interpolate toward usable long-term targets. After an earlier target year, annual growth is capped at 15% through FY2030.
Without a target or cited backlog, growth up to 14% holds its latest rate. At 16% and above, growth fades toward 3%, retaining 80% of the gap each year. Between these bounds, annual growth blends the hold and fade paths linearly. Declines keep their latest rate; existing target and backlog paths are preserved.
For businesses growing above 5%, FY2031 growth is the lower of FY2030 growth and 18%. It then falls in a straight line to 5% in FY2034. Slower-growing and shrinking businesses keep their existing rate.
For continuing cash-flow businesses, terminal value represents the cash beyond the explicit forecast. The standard growth assumption is 3% for growth businesses and 2% for mature businesses. Declining businesses keep their decline, with a floor of -3%.
Terminal growth must stay at least 4 percentage points below the relevant discount rate. Capacity models include ongoing equipment replacement. Projects and REITs use the special rules above.
For operating businesses, the FY2030 value comes from FY2031–FY2034 cash flows and the terminal value, discounted to FY2030. We do not choose an exit multiple to produce a desired share price.
Before publication
A precise-looking output is only useful if the inputs and the bridge to shareholders hold together.
Some checks produce warnings that need an explanation; others prevent publication. A published result can still carry a visible warning.
Accountability
Management can miss its targets. Demand, margins, funding costs and asset prices can change. Our estimates can be too optimistic, and a consistent formula cannot rescue a bad assumption.
Read the conditions each valuation depends on and test the assumptions you disagree with. Fair value is conditional research, not a prediction of the next market price or a guaranteed return.
Our track record records earnings forecasts and model portfolio performance. It is a separate accountability record, not proof that a particular fair value will be reached.
Questions investors ask
The market and our model may assume different growth, margins, investment needs or required returns. A large gap is a reason to inspect those differences. It does not establish that the market is wrong or say when the price will change.
A price target is someone else's conclusion, often for a different horizon. We build value from operating inputs and future cash flows. In a research estimate, cited consensus revenue or profit forecasts can inform those inputs; an analyst's target price does not set our answer.
Management has not given a usable long-term target, so the research desk supplies a sourced inference. Consensus forecasts, company plans, operating assumptions and comparable-company economics can support it. The label distinguishes this from management guidance.
An input Signal Ledger sets itself where the evidence leaves room for judgment, such as a discount rate or the revenue of a business that has not launched yet. It is part of the one published valuation; the report's “Our judgment calls” section gives each call and its reason.
New filings, earnings guidance and material changes in targets or evidence are reasons to rerun the research. Each report shows its research date and sources; an older report is not silently made current by a live share price. Check the date and any freshness warning rather than assuming a fixed daily refresh.
Yes. Open the assumption trial card to change supported sales or discount-rate inputs and see the recalculated value. For business-level detail, open the valuation workspace. Your trial is separate from the published research case.
The research layer gathers filings, guidance and targets and proposes sourced inputs. Deterministic code turns those inputs into cash flows and value using common rules. AI-extracted facts and research judgments can still be wrong; the source trail lets you check them.
No. Growth can require equipment, construction or other spending before it earns cash. If the investment does not earn an adequate return, expansion can reduce value. We also account for corporate costs and financing claims before arriving at value per share.
It explains the value implied by a stated set of assumptions. It does not know your finances, time horizon or tolerance for loss. Use the evidence and sensitivity tools to assess the case; the research is not personalized investment advice.