The most expensive sentence in startups is "let's just build it and see." Validation exists to buy the same learning for 5% of the price — and in 2026, the tools make excuses thin. Here is the ladder, cheapest rung first.
Rung 1: Problem interviews (₹0, one week)
Ten conversations with people who have the problem — about the problem, never your solution. "Walk me through the last time this happened. What did it cost you? What did you try?" You are listening for existing spend and existing workarounds: people cobbling spreadsheets and paying interns are buyers; people saying "interesting" are being polite. The mom-test discipline: stories about the past beat opinions about the future, always.
Rung 2: The smoke test (₹10–40k, two weeks)
A landing page stating the promise (a template site is fine here), a price on it, and ₹10–25k of targeted ads. Measure clicks → signups → "notify me when live" or, braver, a refundable pre-order. Conversion above ~5% visitor-to-signup with real ad traffic is a pulse; crickets at ₹15k spend is the cheapest "no" you will ever buy.
Rung 3: The concierge MVP (₹0–50k, a month)
Deliver the outcome manually before building the machine: match the buyers yourself on WhatsApp, run the reports in Excel, be the algorithm. Ten paying customers served by hand teach more than any prototype — and their money is the only validation currency that never lies. (No-code tools stretch this rung remarkably far in 2026.)
Rung 4: Now build — small
Only after the ladder: a real MVP of the proven core loop (costs here, process here). The ladder's output is your requirements document — you have watched the workflow by hand.
Signals that mean build (and ones that don't)
Build: strangers paid or pre-ordered; manual delivery has a queue; the same problem-story repeats across ten interviews unprompted. Do not build yet: friends said they would use it; a competitor exists so "the market is proven" (their traction, not yours); you are certain but haven't asked anyone; the validation was a survey (surveys measure politeness).
The uncomfortable math
Skipping validation risks ₹8–18L of build on a guess. The full ladder costs under ₹1L and a month — and most ideas fail on rung 1 or 2, which is the ladder working. Passed all four rungs? Bring us the evidence — it makes the build faster, cheaper, and scoped around proven demand.
Frequently asked questions
How do I validate a startup idea without building it?
Climb the ladder: ten problem interviews listening for existing spend and workarounds, a landing-page smoke test with ₹10–25k of ads measuring signup conversion, then a concierge MVP delivering the outcome manually to paying customers. Build only after money changes hands.
What validation signals actually mean 'build it'?
Strangers paid or pre-ordered, your manual delivery has a waiting queue, and the same problem-story repeats unprompted across interviews. Friends' enthusiasm, competitor existence, and survey results are politeness, not validation.
How much does startup validation cost?
The full ladder runs under ₹1 lakh and about a month: interviews are free, smoke tests cost ₹10–40k including ads, concierge MVPs ₹0–50k with no-code tools. Compare against ₹8–18 lakh of unvalidated build.
What is a concierge MVP?
Delivering your product's outcome manually before building the machine — matching buyers on WhatsApp, running reports in Excel, being the algorithm yourself. Ten hand-served paying customers teach more than any prototype and pre-write your requirements.