Why Indian Founders Get Their Market Size Wrong — And What to Do About It

At its peak in 2022, BYJU's carried a $22 billion valuation — a number defended with one core calculation: India had 250 million school-age children, supplemental tutoring was an established multi-thousand-crore category, and digital platforms were going to claim a dominant share of it. Investors agreed. Byju Raveendran, the company's Bengaluru-based founder, raised over ₹40,000 crore across the company's lifetime — one of the largest funding accumulations in Indian startup history. By mid-2024, BYJU's had entered insolvency proceedings. The market it described was real. The market that would pay what BYJU's charged — ₹20,000 to ₹60,000 per year per student, sustained for several years — was a fraction of the figure on the pitch deck.

That gap is not a BYJU's problem. It is India's most repeatable market sizing mistake — specific, structural, and almost never discussed openly in founder circles. The error is straightforward: most founders count everyone with the problem as a potential customer, then apply a percentage to an industry report total. The result is a TAM that is 5 to 20 times larger than the population that will actually pay. The Household Consumer Expenditure Survey released by the Ministry of Statistics and Programme Implementation in February 2024 — the first of its kind in 11 years — gives every Indian founder a free, survey-grade way to find the honest number. Most founders have never opened it.

Why Indian Pitch Deck TAMs Bear No Relation to Reality

The incentive to inflate is structural. Venture capital and angel investors need large addressable markets to justify risk and return math. A ₹400 crore TAM can support a sound business but will not produce a fund-returning exit. So top-down market sizing has become a ritual: locate the largest industry number in a consultancy report, apply a 1-5% capture assumption, and label it the 5-year target. The number passes the filter at a pitch meeting without needing to reflect actual consumer purchasing power.

The near-term cost is low — a credulous investor, a round closed. The delayed cost is significant. Founders hire teams, build infrastructure, and raise subsequent rounds all premised on a market that does not exist at the price point they assumed. Analysis shared by Deutsche Consulting in 2025 found that 11,223 Indian startups shut down in just the first ten months of 2025, a 30% increase over all closures in 2024. Post-mortem research consistently finds that 36-42% of those failures built something the market did not want at the price required. That is a market sizing failure that was never interrogated — often because no one asked the foundational question: of the people who have this problem, how many can actually afford your solution?

The cleanest counterexample is one city away from the company that failed most publicly.

What India's First Household Survey in 11 Years Actually Shows

MoSPI conducted the Household Consumer Expenditure Survey 2022-23 between August 2022 and July 2023, gathering data from 2,61,746 households across rural and urban India. Results were released in February 2024. The previous edition was 2011-12. That 11-year gap means the entire first wave of Indian consumer internet companies was sized against household spending data from a different economic reality entirely.

The headline finding from the MoSPI HCES 2022-23 report: average monthly per capita consumption expenditure is ₹3,773 in rural India and ₹6,459 in urban India. Non-food items now account for about 53% of rural household spending and 60% of urban. But the average obscures the distribution that matters for market sizing.

In urban India, households at the 10th percentile spend ₹2,810 per month — total, on everything including food, rent, transport, and communication. Households at the 90th percentile spend ₹11,084 per month — a 4x gap, per the same MoSPI data. In rural India, the spread runs from ₹1,911 at the 10th percentile to ₹6,043 at the 90th percentile. For a consumer product priced above ₹300 per month, or a one-time purchase above ₹2,000, the realistic paying population is the top two income quintiles in urban India — roughly 50-60 million urban households. Not 450 million urban residents, and certainly not 1.4 billion Indians.

Urban India's bottom income decile spends ₹2,810 per month on everything. If your subscription costs ₹500, you have already claimed 18% of their total monthly budget before they buy food. That is not your customer — that is a segment you are pricing out before you launch.

The implication for market sizing is direct. Any consumer startup that multiplies an Indian population count by a unit price needs to layer in the HCES quintile data first. The question is not how many Indians have the problem. It is how many Indian households have enough discretionary spending after food, rent, and transport to afford your price. The HCES answers that, by quintile, for rural and urban separately, at state level. Nothing else free in India provides this kind of granularity.

Top-Down vs Bottom-Up: The Same Idea, Three Very Different Numbers

Run the same category through two approaches and you see the pattern clearly.

Vernacular online education

Top-down: India has roughly 2 million JEE and NEET aspirants per exam cycle. At ₹15,000 per year average spend, the market is ₹3,000 crore. Bottom-up using HCES: JEE and NEET preparation is concentrated in households spending above ₹4,000 per person per month — a threshold that excludes roughly 60-65% of Indian households at their current income levels. Of the remaining pool, active exam-prep families number around 2-3 million. But approximately 70% are in Tier 2 and Tier 3 cities, where monthly household income puts ₹15,000 per year for a single online course well outside comfortable reach. The realistic paying pool for a platform charging ₹15,000 per year sits closer to 4-6 lakh families. At ₹5,000 per year — the price that Tier 2 cities can genuinely absorb — the honest market is ₹200-300 crore. Still fundable. Just 10x smaller than the slide says.

B2B SaaS for restaurant management

Top-down: India has 7.5 million food service establishments. At ₹1,200 per month per outlet, the market exceeds ₹1 lakh crore per year. Bottom-up using Udyam: the MSME Ministry's Udyam Registration Portal had 7.83 crore registered enterprises as of February 2026, growing from 4.12 crore in March 2024, per the Ministry's official data. The food and beverage segment accounts for roughly 8-10% of registrations — about 60-70 lakh establishments. Over 85% of those are single-location micro-businesses with monthly revenues under ₹1 lakh, where a ₹1,200-per-month SaaS subscription equals 1-1.5% of gross revenue with no margin to absorb it. The addressable market for restaurant software at that price is chains and multi-location operators — roughly 2-4 lakh establishments. A real bottom-up SAM of ₹2,400-4,800 crore. Fundable, honest, and around 20 times smaller than the headline.

Premium D2C skincare

The skincare calculation follows the same logic. India's cosmetics market is commonly cited at ₹35,000-40,000 crore, so a 1% share looks like ₹350-400 crore in a deck. But the HCES 2022-23 data shows that urban households in the bottom two income quintiles spend less than ₹200 per month on all personal care combined — soap, shampoo, toothpaste, face wash, and anything else in the bathroom. A D2C brand priced at ₹500-800 per SKU is genuinely targeting the top 20-25% of urban households. Once you factor in brand awareness curves and the distribution cost of reaching that segment reliably, the realistic SAM is a number that still supports a real business — but not the one that appears in the top-down calculation.

The economics look different depending on the segment you target. The hyperlocal D2C skincare opportunity is built specifically around the urban Tier 1 household profile — where the HCES top quintile confirms genuine skincare budget — rather than a broad national count.

Five Free Government Databases That Replace the ₹3 Lakh Market Report

Most founders pay market research agencies ₹2-5 lakh for reports that pull their headline numbers from a small set of government databases and add narrative. The databases are all free and publicly accessible:

• HCES 2022-23 (MoSPI) — At mospi.gov.in, the factsheet and detailed tables give household expenditure by state, by income quintile, and by food versus non-food category. This is your primary source for any consumer product market sizing. Released in February 2024 after an 11-year gap; the follow-on HCES 2023-24 was released in December 2024. Ignore any market report that does not reference this data for India consumer sizing — it is working from 2011-era numbers.
• PLFS 2023-24 (Periodic Labour Force Survey) — Also from MoSPI, available at data.gov.in. The PLFS gives employment status, occupation type, and earnings by state, sector, and rural-urban split. If your product is for workers of any kind — gig platform sellers, MSME owners, salaried employees — this is the denominator for your SAM. It is updated annually.
• Udyam Registration Data — dashboard.msme.gov.in shows district-level counts of registered MSME enterprises by type: micro, small, or medium. As of February 2026, 7.83 crore enterprises are registered, per MSME Ministry data. For any B2B tool aimed at small businesses, this is the free alternative to a ₹50,000 commercial database, with the added advantage of being filterable by district and enterprise size.
• MCA21 Company Master Data — The Ministry of Corporate Affairs hosts company master data at data.gov.in: corporate identification numbers, registration dates, sector classification, paid-up capital, and current status. In FY 2024-25, 1.12 lakh new companies registered, per the Ministry of Corporate Affairs. For competitive density mapping and understanding which sub-sectors are newly active, this is available at no cost.
• ONDC Open Data — opendata.ondc.org publishes actual transaction volumes, product categories, and geographic distribution from the government-backed open commerce network. By end of 2024, the network was processing 15 million-plus orders per month, with over 70% of its seller base comprising small and medium businesses, per ONDC transaction reports. For any product going into e-commerce, this is real demand data rather than projection.

A ₹3 lakh market research report is usually a formatted version of data you can access for free from MoSPI, the MSME dashboard, and the Ministry of Corporate Affairs. The agency adds narrative. You can add your own. What you cannot fabricate is the underlying household-level survey data — and the government publishes that.

For founders building a B2B tool targeting the 7.83 crore Udyam-registered MSMEs, the fintech platform for kirana retailers is a useful reference for how unit economics work when your customer is a micro-business with sub-₹5 lakh annual revenue.

The Lesson PhysicsWallah Built In From Day One

Alakh Pandey started teaching physics on YouTube from Prayagraj in 2016. The coaching institutes he had worked for charged ₹1 lakh to ₹1.5 lakh per year — fees calibrated for families in large cities with two earning members and disposable income well above the HCES urban average. He built his platform in Noida and priced it at ₹500 to ₹1,500 per year, specifically designed for households whose monthly per capita spending put them in the ₹4,000-6,000 range — people with real ambition for their children and a hard ceiling on what they could spend annually on education technology.

Prateek Maheshwari, co-founder of PhysicsWallah, had built a separate e-learning application and merged it with Pandey's YouTube community when they formally incorporated the company in Noida in 2020. They targeted 2-3 million JEE and NEET aspirants in Tier 2 and Tier 3 cities who could not afford residential coaching. That was not a residual market — it was the majority of India's competitive exam aspirants, the slice that the ₹1.5 lakh per year coaching model had always priced out.

By September 2024, PhysicsWallah was valued at $2.8 billion, per BusinessToday reporting at the time. In the first nine months of 2024 alone, the company raised $210 million. PhysicsWallah now reaches 46 million students and operates at 98% of India's postal codes. Meanwhile, BYJU's — which priced its app at ₹20,000-60,000 per year and tried to sell it across India's income distribution — entered insolvency. The difference was not technology. Both companies used video lectures. The difference was a founder who sized his market around actual income levels rather than population counts.

The same income-aware approach applies to vernacular skill-based edtech platforms where the HCES data points to a large working-adult segment with regional-language preferences and ₹2,000-5,000 per course budget — a real market that premium English-medium competitors have systematically underpriced.

How to Build Your India Market Size in Four Hours

A credible market size document for an Indian startup takes four hours and costs nothing, using government data:

1. Open the HCES 2022-23 factsheet at mospi.gov.in. Find the household expenditure category closest to your product. Note the monthly per capita spend broken by quintile for urban and rural separately. Your TAM denominator is the population in the top two or three income quintiles — depending on your price point — multiplied by annual spend in your category. Do not use the country-average. Use the quintile that matches the household profile your pricing actually requires.
2. If your product is B2B, open dashboard.msme.gov.in and filter by state, district, and enterprise type. Count registered businesses in your sector and target geography. Cross-reference with the MCA21 company master data at data.gov.in to understand how many are micro versus small versus medium — the revenue ceiling of a micro business is ₹1 crore per year, which sets a hard limit on what you can charge.
3. For e-commerce or physical-goods products, check opendata.ondc.org. Real transaction volume by category and city tells you whether demand exists before you commit to a supply chain. A category conspicuously absent from ONDC data despite being described as 'fast-growing' in a 2022 industry report deserves investigation before you build.
4. Run Google Trends filtered to India and to your three most important target states, set to a 5-year window. A category in structural search decline while still appearing in bullish pitch decks is a signal worth understanding. Demand that grew sharply in 2020-21 and has since flattened or declined is the ghost of pandemic-era TAM estimates that never reflected steady-state India.

Combine these four inputs and you will have a market size figure that is smaller than the top-down number. It will also be the number you can defend when a Series A investor asks where it comes from — a question that India-focused investors, who lived through BYJU's and Unacademy's implosions, are increasingly asking.

The market research framework for finding demand signals in free Indian data sources is covered step by step in our market research playbook, including Google Trends filtering by state, MCA21 queries, and ONDC transaction category lookups.

For the demand-side test that confirms someone will actually pay before you build — the UPI signal rather than the survey — see the business idea validation guide which covers the WhatsApp MVP and fake-door approaches used by Dunzo and Mamaearth at the pre-seed stage.

The Cost of Getting This Wrong

BYJU's, founded by Byju Raveendran in Bengaluru, entered insolvency proceedings in 2024 after raising over ₹40,000 crore across its lifetime. That was not a surprise to anyone who had looked closely at its customer-level unit economics. The company had built its valuation on a TAM that counted school-age children, not families with the income and willingness to sustain ₹20,000-60,000 per year in education tech subscriptions. When the pandemic-era willingness to experiment wore off, the income distribution of actual Indian households set the ceiling on what BYJU's could recover.

The HCES 2022-23 data was not available when BYJU's raised its earliest large rounds. It is available now, published by MoSPI in February 2024. Any founder who prices a consumer product or builds a B2B tool for Indian MSMEs in 2025 or 2026 without referencing it is choosing to work with less information than the situation allows. That is not a neutral choice. It carries a cost that has been paid visibly, at scale, over the last two years.

The honest market size is smaller than the pitch deck number. It is also the right size to build from. PhysicsWallah sized its market at the slice of India that could genuinely pay ₹500-1,500 per year and built one of the most capital-efficient edtech companies in the world as a result. The HCES data is where that kind of clarity begins.

Meesho is another case of a company that correctly sized Tier 2 and Tier 3 India at income-appropriate price points — its model is examined in our Meesho social commerce deep dive, which details the specific pricing mechanics and supplier economics.

Quick commerce is a sector where the gap between projected TAM and real addressable density became apparent within two years of launch — our Zepto case study looks at how hyperlocal dark store density, not national scale, is the correct unit of market analysis for 10-minute delivery.

For founders sizing a SaaS business targeting food service operators, the restaurant and cloud kitchen management SaaS idea includes revenue benchmarks built from the multi-location operator segment — the 2-4 lakh establishments that can actually pay ₹1,200 per month — rather than the 7.5 million total outlets figure.

Last updated: July 2026