EV/revenue calculator

Because applying an EV/revenue multiple to a profitable, mature business would silently produce a meaningless number — and silent wrong answers are worse than refusals. Every other lens in the suite gates softly (it runs but flags required:false); this one is different. If the archetype is anything but pre_profit, the calculator returns EXCLUDED rather than computing. For a pre-profit name it is the primary lens; for a mature compounder or a financial, the FCFF DCF and EPS-based DCFs are correct, and this model declines to run at all. That refusal is the design, not a limitation.

Roughly 20–25% mature EBITDA margins and a low-single-digit yield to capital providers — i.e. a high-quality, profitable, growth-decelerating software business at scale. 8× is the calibration anchor when the archetype hint supplies nothing. If you do not believe the company can reach ~20% EBITDA margins, the 8× anchor is wrong — try 4–5×. If you think it will be a premium SaaS leader (Adobe-like), supply 12–15× via the archetype calibration. The hard clamp is [1, 30]: the 30× ceiling blocks speculative-bubble multiples (peak-2021 software ran 30–50× and proved unsustainable); the 1× floor blocks a distressed multiple from collapsing the model.

Because pre-profit valuations genuinely carry more uncertainty, and the range should say so. Instead of a flat band around the deterministic mid, this calculator independently stresses both axes that drive the answer: the low scenario takes growth −3pp/yr AND the multiple ×0.75; the high takes growth +3pp/yr AND ×1.25. Small changes in the terminal multiple compound through five years of growth into multi-× differences in fair value. The three results are sorted ascending afterwards — so even if a stressed growth path floors equity at zero and the "low" mathematically exceeds the "mid", the labels stay semantically correct (low ≤ mid ≤ high).

In priority order. First: an LLM-supplied explicit path — up to five rates the analysis pipeline hands over (e.g. 55%, 40%, 30%, 22%, 15%). A percent-vs-decimal heuristic means it tolerates either [55, 40, …] or [0.55, 0.40, …] — any value with absolute size above 1.5 is divided by 100. Second, if no explicit path exists: a deterministic fade from a single starting rate (default 15%) decelerating 8pp/year toward a 3% terminal floor. The fade is intentionally conservative; genuinely high-growth names should carry an LLM path that fades more slowly. Every rate is clamped per-year to [−50%, max growth cap (default 100%)], so even a supplied 200% rate is capped — preventing a $500B fair value on a $50M-revenue company.

equity = present EV − net debt. If net debt exceeds the present enterprise value the deterministic projection supports, equity is negative and per-share fair value floors at max(0, equity)/shares = 0. That is informative, not an error: the debt load is larger than the chosen revenue projection and multiple can support. Either the multiple is too low (raise it if the margin story is credible), the growth path is too conservative, or the company has a genuine capital-structure problem. Net debt matters a lot here — many 2023+ SaaS names took on real debt — so the audit always shows the figure used, and reliability drops 0.05 if it was derived (debt − cash) rather than supplied explicitly.