Two methods, because one is never enough.
Every market sizing exercise is a triangulation problem. Top-down analysis starts from a population — total spend, total addressable users, total industry revenue — and filters down to your segment. Bottom-up analysis starts from a unit — one customer’s annual spend, one transaction’s economics — and multiplies up.
Both methods are flawed in opposite directions. Top-down inherits the assumptions of whoever published the population number. Bottom-up inherits the optimism of whoever sized the unit. Used alone, either is fragile. Used together, the gap between them tells you something neither method tells you on its own: where your assumptions are weakest.
Sizoru runs both methods on every report. We don’t pick a winner. We surface the convergence — the band where both methods agree — and we flag the divergence, with reasoning, when they don’t.
A market sized by one method is an opinion. A market sized by two convergent methods is a defensible thesis.
That is the entire premise of this tool. Everything below is how we operationalize it.
Method 1 — Top-down: from the universe inward.
Top-down begins with the largest defensible number — total industry revenue, total addressable population, total category spend — and applies sequential filters until what remains is your addressable market.
Worked example · B2B SaaS startup selling to Indian SMEs
| Step | Filter | Source tier | Resulting figure |
|---|---|---|---|
| 1 | Total registered MSMEs in India | Tier 1 — MoSPI, Udyam Registration | ~6.3 cr |
| 2 | Filter to digitally-active MSMEs | Tier 2 — NASSCOM SME Digital 2025 | ~1.8 cr |
| 3 | Filter to your sub-vertical (F&B retail) | Tier 1 + 2 cross-reference | ~22 lakh |
| 4 | Filter to those willing to pay for SaaS | Tier 2 — industry surveys | ~6 lakh |
| 5 | Multiply by ARPU benchmark | Tier 2 — comparable companies | ₹1,200 cr TAM |
This is the top-down candidate number. We do not stop here.
Method 2 — Bottom-up: from the unit outward.
Bottom-up begins with what one customer is worth — annual contract value, transaction frequency, lifetime spend — and multiplies by the realistically reachable customer count.
Same B2B SaaS startup · sized from the unit
| Step | Component | Source tier | Value |
|---|---|---|---|
| 1 | Average ARPU per SME customer | Tier 2 — comparable Indian SaaS | ₹18,000 / yr |
| 2 | Annual retention rate | Tier 2 — SaaS Capital benchmarks | 78% |
| 3 | Realistic customer ceiling (5 yr) | Bottom-up build | ~3.5 lakh |
| 4 | Average customer lifetime | Derived from retention | 4.5 years |
| 5 | Bottom-up TAM | Computed | ₹2,835 cr |
Top-down said ₹1,200 cr. Bottom-up says ₹2,835 cr. Now we have the interesting part.
The convergence test.
When two methods agree, we lock the number. When they don’t, we tell you why.
A top-down number of ₹1,200 cr and a bottom-up number of ₹2,835 cr — in the example above — represent a 136% gap. That gap is not a problem. It is information. Sizoru applies a tiered convergence rule:
Both methods agree. We ship a single TAM with high confidence.
Methods agree on the ballpark. We ship a TAM range with both endpoints fully cited.
One assumption is breaking. We ship both numbers, plus an analytical breakdown of which assumption diverges and why.
In the worked example, the 136% gap suggests that either (a) the top-down filter for “digitally active MSMEs willing to pay” is too aggressive — meaning bottom-up is closer to truth — or (b) the bottom-up customer ceiling of 3.5 lakh is unrealistically high — meaning top-down is closer.
We don’t pretend to know which. We surface both, explain the divergence, and let the founder defend their thesis with both methods visible.
A VC who sees a converged number has confidence. A VC who sees a flagged divergence with honest reasoning has more confidence — because they know the founder didn’t paper over the gap.
India is not one market.
So we don’t size it as one.
Indian market sizing is uniquely broken. Most reports — including premium ones from global firms — treat India as a single 1.4 billion consumer block, then apply uniform filters. The result is a number that is technically defensible and operationally meaningless.
Sizoru applies the India 1 / 2 / 3 framework on every report involving Indian geography. It’s a segmentation lens, not a number-cruncher — but it changes the SAM by 40 – 70% in most cases.
India 1
~120 million people
- Per-capita comparable to Mexico / Brazil
- Pays in USD-equivalent for premium goods
- Tier-1 metro, English-comfortable
Most premium SaaS, fintech, and D2C lives here
India 2
~300 million people
- Aspirational middle, mostly digital-first
- Pays in INR for value-driven SKUs
- Tier-2 / 3 cities, vernacular-first
Most quick-commerce, gaming, and EdTech lives here
India 3
~1 billion people
- Largely informal economy, low discretionary
- Limited monetization beyond utility
- Rural + urban informal
Most subsidized and utility-led products live here
A SaaS startup targeting India 1 has a SAM around ~120 million addressable users — not 1.4 billion. A vernacular EdTech startup targeting India 2 has a different ceiling and a different ARPU curve. Conflating the three produces market sizes that look impressive in a deck and collapse under investor scrutiny.
Every Sizoru report involving Indian geography asks: which of the three are you actually selling to? The SAM is filtered to that segment. The investor sees a TAM that respects the country it’s drawn from.
Three tiers of sources.
Marked clearly. Used differently.
Not all sources are equal. A government statistic and a Mordor Intelligence projection are not interchangeable, even when they appear in the same paragraph. Sizoru classifies every cited source into one of three tiers, marks each with a visible badge in the report, and weights them accordingly in the math.
Examples: RBI, SEBI, MoSPI, NSO, Census of India, NPCI, regulator filings, World Bank, IMF.
Used: Used as primary inputs without qualification. Cited as ground truth.
Examples: McKinsey, BCG, Bain, KPMG industry papers, NASSCOM, FICCI / CII white papers, named comparable filings, peer-reviewed research.
Used: Used as supporting inputs. Cited with the publishing firm visible.
Examples: Grand View Research, Allied Market Research, Mordor Intelligence, MarketsandMarkets, generic third-party PR.
Used: Used only as triangulation, never as a sole source. Marked with a ⚠ caution badge in the report. A number sourced exclusively from Tier 3 is flagged for what it is: a syndicated estimate, not a measurement.
Every figure in every Sizoru report carries its tier badge in the source footnote. If a number you’re about to put in your investor deck is Tier 3, you’ll know it before the VC does.
Every TAM → SAM → SOM step is shown.
No filter chain is hidden.
The standard market-sizing handoff looks like this: “TAM is ₹10,000 cr, SAM is ₹2,000 cr, SOM is ₹200 cr.” Every founder has put a slide like this in a deck. Every VC has nodded politely and stopped trusting the deck.
The reason: the filtering is invisible. Why did 80% of the TAM disappear into “SAM”? Why is SOM 10% of SAM and not 5% or 20%? Without the filter chain visible, the numbers are decorative.
TAM ₹10,000 cr
─ Filter 1 · Geography (India only) → −25%
─ Filter 2 · Segment (B2B only) → −40%
─ Filter 3 · Sub-vertical (F&B retail) → −55%
SAM ₹2,025 cr
─ Filter 4 · Channel reachability (3 yr) → −60%
─ Filter 5 · Pricing fit (within ARPU) → −50%
─ Filter 6 · Capacity (your team’s reach) → −75%
SOM ₹202 cr
Every Sizoru report ships a filter chain like the one above — with each filter labeled, each percentage shown, each rationale footnoted. A VC reading the report sees not just the numbers but the thinking behind the numbers. Pushback becomes specific (“I disagree with Filter 4”) instead of dismissive (“the numbers seem high”).
Specific pushback is debatable. Dismissive pushback is fatal. We optimize for the former.
Three scenarios.
Because no thesis survives a single number.
Every Sizoru SOM is delivered as three scenarios — bull, base, bear — with the assumption table that produces each. The base case is the realistic 5-year projection. The bull case is the upside if execution and market timing both hit. The bear case is what happens if growth assumptions compress by 30 – 50%.
| Variable | Bear | Base | Bull |
|---|---|---|---|
| 5-yr CAGR | 12% | 24% | 38% |
| CAC growth (yoy) | +50% | flat | −20% |
| Net revenue retention | 95% | 110% | 130% |
| 5-yr SOM | ₹85 cr | ₹202 cr | ₹420 cr |
Founders who walk into a room with three scenarios out-credentialize founders with one. Bear-case rigor is the most underrated trust signal in venture pitching.
What Sizoru is not.
We don’t predict the future.
Sizoru sizes markets as they are and projects under stated assumptions. The future is a probability distribution. We give you the distribution; we do not collapse it for you.
We don’t replace primary research.
If you need a 200-respondent customer survey, you need a research firm. Sizoru is a top-down + bottom-up sizing tool; it is not a substitute for talking to your buyer.
We don’t size markets that don’t exist yet.
Sizoru requires at least one comparable industry or one defensible analogue. A market with zero comparables is a market we’ll tell you we can’t size — and that itself is useful information.
We don’t promise the AI is always right.
Every Sizoru report is generated, then validated against named-source benchmarks. When validation flags an inconsistency, we surface it. We do not pretend the model is infallible.
We don’t sell the report alone as a thesis.
The report is evidence. Your conviction is the thesis. We will never claim that downloading our .docx is a substitute for understanding your market.
Who built this methodology.
Sizoru is a product of Stolvix, an Indian technology company building investor-grade analytical tools for founders. Sizoru was created to bring institutional- grade market sizing methodology — typically gated behind ₹3 – 5 lakh consulting engagements — to the founders, consultants, and investors who need it weekly.
The dual-method framework, India 1/2/3 segmentation, and tier-rated source provenance documented above are codified in the Sizoru Market Sizing Framework v1.0.0, the proprietary methodology of Stolvix and the canonical reference applied to every report this product ships.
Questions on methodology? pot@1xl.com
See it on a real market.
The methodology above produces a polished .docx report, .pptx investor deck, and print-ready .pdf on any market you describe. ₹14,999 per report. Three free revisions in 30 days. Pass investor review or we redo it.