Cirrus Data: A DCF That Is Really a Terminal-Margin Bet
You are valuing Cirrus Data (CRDS), a subscription software company, using a five-year DCF. The model produces $96 per share. Terminal value is 78% of enterprise value. The terminal year assumes a 32% EBIT margin, versus 9% today and a company best of 14%. WACC is 9.5%, built from a 4.2% risk-free rate, a 5.5% equity risk premium, a beta of 1.20, a 20% debt weight, and a 4.0% after-tax cost of debt. Terminal growth is 3.0%. Peer median beta is 1.45. Comparable companies trade at 6.2x 2026E revenue; your target implies 7.4x. A portfolio manager asks why she should pay 7.4x for CRDS when the group is at 6.2x.
1. Which assumption actually drives the $96 value, and is it defensible?
Read the worked answer
The five-year forecast is decoration; the value is the terminal assumption. With 78% of enterprise value beyond year five, the number that matters is the 32% terminal EBIT margin - more than triple today's 9% and more than double the best margin the company has ever produced.
That is not automatically wrong; young subscription businesses do scale. But it is a claim that requires a demonstrated mechanism: gross margin stability as hosting costs scale, sales and marketing falling from roughly half of revenue toward a normalized level, and R&D leverage. Ask for the terminal-year cost stack, line by line, and compare each line to the best operator in the peer group. If the 32% requires sales efficiency above anyone in the sector, it is an assumption, not a forecast.
The secondary problem is the discount rate. A beta of 1.20 against a peer median of 1.45 lowers the cost of equity and inflates every terminal dollar. In a high-multiple, high-volatility software name, a below-peer beta needs a justification - a longer trading history, a genuinely lower covariance - not just a regression window that happened to produce it.
2. Quantify the sensitivity: what does a 50bp change in terminal growth, and a peer-median beta, do to the value? Compare that with the effect of a lower terminal margin.
Read the worked answer
The terminal value multiple is 1/(WACC − g) = 1/(0.095 − 0.030) = 15.4x terminal free cash flow.
- Growth to 2.5%: 1/0.070 = 14.3x, terminal value down about 7%, enterprise value down about 5.5% (0.78 × 7%).
- Peer beta of 1.45: cost of equity = 4.2 + 1.45 × 5.5 = 12.2%; WACC = 0.8 × 12.2 + 0.2 × 4.0 = 10.5%. Terminal multiple = 1/0.075 = 13.3x, terminal value down about 13%, enterprise value down about 10%. The per-share hit is larger than 10% because the enterprise value falls while net debt is fixed.
- Terminal margin to 20% from 32%: terminal NOPAT, and therefore terminal value, falls roughly 37%, so enterprise value falls about 29% (0.78 × 37%).
The ranking is the point. A 100bp swing in WACC or a 50bp swing in growth moves the answer by single digits to low double digits; the margin assumption moves it by nearly a third. Diligence should be spent proving the margin path, not tuning the discount rate to three decimal places. Present the value as a range driven by terminal margin, and disclose what margin the current price implies.
3. How would you reconcile the DCF with the 7.4x implied multiple against a 6.2x peer median?
Read the worked answer
A 19% premium requires a reason that survives being written down.
First, test the comp set before defending the premium. If the median includes slower-growth, lower-retention or SMB-focused names, the median is not the right benchmark and the premium may be an artifact of a bad peer group.
Second, if the set is sound, decompose the premium into growth, retention, and margin. A growth-adjusted multiple or a Rule of 40 comparison (growth plus FCF margin) will show quickly whether CRDS deserves 7.4x. Twenty-five percent growth at a 5% FCF margin does not obviously beat fifteen percent growth at a 10% margin.
Third, do not lean on EV/Revenue alone. Revenue multiples ignore margin and capital structure; pair it with EV/gross profit and EV/unlevered FCF so the margin assumption baked into the DCF is visible in the multiple.
Finally, accept the possibility that the DCF is wrong. A DCF calibrated to land on a premium the analyst already wanted is reverse engineering. If the honest terminal margin produces a value implying 5.5x, say so and let the recommendation follow - either the stock is expensive or the peer group is mispriced, and the report should pick one and defend it.