Why put up with EV/Sales here, when Phase 3 taught you to distrust it? Because growth-stage software earnings are deliberately suppressed, as the subscription machine showed, so E-based multiples come out meaningless or infinite. Revenue, meanwhile, is (a) real, (b) high-gross-margin, and (c) annuity-like. A rupee of SaaS revenue isn't a rupee of construction revenue. Still, EV/Sales is a loaded gun with two safety catches, and you never fire it without both.
How do you adjust EV/Sales for growth and margin?
Safety catch 1: growth adjustment. A revenue multiple is a compressed DCF whose main hidden variable is growth. Don't ever compare two EV/Sales numbers without dividing by growth. 12× at 40% growth (0.30 per point) can be cheaper than 5× at 10% (0.50 per point). Market context moves the whole grid with rates and sentiment. As a rough guide, as of mid-2026, median public SaaS trades around 5–6× NTM revenue, with the >25%-growth cohort near 9–12×. Verify that against a live comp sheet. Jamin Ball's weekly Clouded Judgement tables, Meritech's public SaaS comps, or the BVP Nasdaq Emerging Cloud Index page will do, and all are free. Rule-of-40-adjusted comparisons work the same way. Plot EV/Sales against Rule-of-40 score across the comp set and price off the regression line instead of the raw multiple.
Safety catch 2: the margin-maturity bridge. A revenue multiple can only be justified by the earnings power it implies at maturity. Force every EV/Sales through this four-step bridge:
Step 1 Project revenue forward at a defensible growth path
Step 2 Apply a steady-state FCF margin (what does THIS model earn when S&M normalizes?)
Step 3 Apply a mature-company multiple to that FCF
Step 4 Discount back and compare to today's EVAdvertisement
Worked example: a SaaS company priced at 12.6× sales
Try it with numbers. A SaaS company with ₹1,000 cr revenue is growing 30% and priced at ₹12,600 cr EV (12.6× sales). Five years at 30% takes revenue to 1,000 × 1.30⁵ = ₹3,713 cr. For a good sticky model, take a steady-state FCF margin of 30%. Best-in-class mature software prints 30–40%+ and mediocre sticky software 15–25%, so pick yours from gross margin minus normalized opex. Hope is no basis for it. That gives FCF of ₹1,114 cr. A durable low-teens grower deserves a mature multiple of 20× FCF, so year-5 EV is ₹22,278 cr. Discount at 12% (1.12⁵ = 1.762) and you get ₹12,641 cr today. The 12.6× price is fair, then, if you believe in 30% growth for five years and a 30% terminal margin. Now you know what you're underwriting, and that's what the exercise is for.
Run it in reverse, the reverse-DCF reflex. Suppose the market pays 20× (EV ₹20,000 cr) for the same company. Required year-5 EV = 20,000 × 1.762 = 35,247. That needs FCF of 1,762, which needs revenue of ₹5,874 cr, which implies growth of (5.874)^(1/5) − 1 ≈ 42.5% a year for five years. Actual growth is 30% and slowing. You don't need a view on the stock to see what the price assumes. You can also ask whether any company of this size in this category has ever achieved it, which is the base-rate question.
EV/ARR or EV/Sales, and what happens as growth slows?
EV/ARR vs EV/Sales. EV/ARR uses the forward-looking run-rate (closing or next-quarter ARR) instead of trailing revenue. For fast growers it's cleaner, since trailing revenue understates the current book. For VyapaarDesk, EV/Sales on 480 and EV/ARR on 512 differ by ~7%. Use EV/ARR when the ARR is clean, meaning rungs 1–3 of the quality ladder, cross-checked. Use EV/Sales when it isn't, or when ARR isn't disclosed. Never mix the two across a comp table.
The last piece is the margin-for-growth swap over a company's life. As growth slows, S&M intensity falls and margins rise, so the multiple compresses much more slowly than growth does. Take a company that goes from 30% growth/0% FCF to 10% growth/30% FCF. It hasn't become worthless, just a different animal, priced on a different leg of Rule of 40. Worked example 1 (Salesforce-style) shows this lifecycle with numbers. Failing to see it coming is how growth investors overpay at the top, by extrapolating growth. It's also how value investors miss great entries at the transition, by screening on a trailing P/E that's about to inflect.