Ask what your startup is worth and you'll get theatrical answers — spreadsheets pretending to precision, founders quoting hopes, investors quoting portfolios. The truth practitioners know: early-stage valuation is negotiated storytelling anchored by a handful of methods, each useful mainly for keeping the negotiation honest. Here are the methods actually used, what each is good for, and India's peculiar extra layer — the statutory valuation reports the law demands regardless of the negotiation.
Comparables: the market's actual method
What did similar companies raise at? Revenue multiples by sector and stage (Indian SaaS at x-times ARR, consumer brands at y-times revenue, per current market temperature), adjusted for growth rate, margins and market size. This is 80% of real-world early-stage pricing — investors triangulate your round against their last ten deals. Founder's use: build your own comps list (funding announcements, known multiples), know where you sit and why you deserve the top of the band (growth, retention, moat) — arguing multiples beats arguing dreams.
The VC method: working backward from exits
- Investor logic: 'plausible exit value ÷ required return = today's justified price' — a company plausibly worth ₹800 crore in 6 years, needing 10x for fund math, prices today near ₹80 crore post-money
- This is why market size dominates pitches: small plausible exits cap today's valuation regardless of execution
- Founder's use: understand the investor's return arithmetic (fund size, ownership targets — most want 10–20% and returns that matter to their fund) and you understand their price rigidity
DCF and scorecards: the supporting cast
DCF (discounting projected cash flows) is theoretically pure and practically decorative before predictable cash exists — five assumptions stacked on hope; it earns relevance for profitable, steady businesses (and features in statutory reports). Scorecard/Berkus-style methods — rating team, product, market, traction against benchmark values — formalise angel intuition for pre-revenue rounds; useful as sanity structure, not gospel. Cost/asset approaches price what building it again would cost — the floor conversation, relevant mainly in acqui-hires and distress.
Across methods, the same levers: growth rate (the multiple's multiplier), retention/repeat (quality of revenue), gross margin (what scale will yield), founder-market fit, and scarcity (two term sheets beat every model ever built). Valuation work, honestly understood, is metric work plus process work.
India's statutory layer: the reports you'll need anyway
- Company-law: issuing shares above face value needs a Registered Valuer's report (Section 62/Rule 13 contexts) supporting the price
- Income-tax: Rule 11UA valuations (merchant banker for DCF under 56-series contexts) — the angel-tax era's machinery, still relevant for transfer pricing of shares (50CA/56(2)(x)) even after 56(2)(viib)'s abolition for most issues
- FEMA: foreign investors' entry/exit prices bracketed by internationally-accepted-methodology valuations (floor on entry, cap on exit for residents buying back)
- The practical point: the negotiated number and the statutory report must be reconcilable — round pricing should be designed with the valuation report, not retrofitted against it
Negotiating with the methods
Use each tool for its job: comps to anchor the band, the VC method to understand their constraints, your driver-model to defend the top of the band, and process (parallel conversations, real deadlines) to create the scarcity no model supplies. And keep the wisdom that outlasts every method: valuation optimises one round; ownership, preferences and the right partner compound across all of them. A slightly lower price from the investor who triples your odds is the cheapest capital on the table.
How Aidwish helps
Aidwish supports pricing conversations end to end — comps and driver models, statutory valuation coordination (Registered Valuer/merchant banker reports), FEMA-compliant pricing for foreign cheques — so the negotiated story and the regulatory paper tell one number.