Skip to content

12. Market Sizing

Market sizing is not an investor slide exercise. It is a way to understand whether the opportunity can support the company you want to build. A market can be too small for venture capital and still be a wonderful business. A market can look huge in reports and still be impossible for you to enter.

Good market sizing helps you answer: if we win our first wedge, what can this become?

The core sizing question is: how many reachable customers have this problem, what can they realistically pay, and what expansion path makes the first wedge worth pursuing?

For founders, market sizing should create discipline, not decoration. The goal is to understand whether the opportunity fits the company you want to build: venture-scale startup, profitable software business, services-led company, community business, marketplace, or cash-flow business.

  • TAM, SAM, and SOM in plain language.
  • Bottom-up versus top-down sizing.
  • First market versus future market.
  • India-specific sizing traps.
  • How to build a useful early market-size estimate.
TermMeaningUseful Question
TAMTotal addressable market.If everyone who could use this bought, how large is the total opportunity?
SAMServiceable available market.Which part can our product and geography realistically serve?
SOMServiceable obtainable market.What share can we plausibly win in the next few years?

The mistake is making TAM the hero. Investors may ask for TAM, but founders should care more about the path from beachhead to expansion. A believable SOM with a strong wedge is often more useful than a giant TAM with no entry plan.

Market sizing should prove different things at different stages.

You are checking whether the problem is worth discovery. A rough estimate is enough. The goal is to avoid spending months on a market that obviously cannot support your ambition.

You are checking whether the first segment is big enough for meaningful learning and early revenue. The exact TAM is less important than the density of reachable customers.

You need to show a credible path from wedge to large opportunity. Investors do not need every number to be perfect, but they need to trust your logic.

You need sharper segmentation, pricing, adoption, churn, gross margin, and channel assumptions. At this stage, bad sizing can create hiring and burn mistakes.

Top-down sizing starts with reports: industry revenue, number of businesses, consumer spend, government data, or analyst estimates. It is useful for orientation, but dangerous when copied blindly.

Example:

“India has X million SMEs. If we charge each Y per month, the market is huge.”

This ignores digital maturity, willingness to pay, buyer access, churn, onboarding cost, support cost, and whether the SME segment is one market at all.

Use top-down sizing only as a rough boundary.

Top-down sizing is most useful when it helps you understand category scale, growth direction, adjacent markets, investor expectations, and whether the opportunity is obviously too small or plausibly large. It is weak for predicting your first customers.

Bottom-up sizing starts with real customer segments.

Formula:

Number of target customers x realistic annual revenue per customer = annual market opportunity

Better version:

Reachable customers x adoption rate x annual contract value x gross margin = practical opportunity

Example:

InputEstimate
Target segmentIndian D2C brands doing 100-1,000 orders/day
Number of reachable companies5,000
Likely adoption over 5 years20%
Annual revenue per customerRs 1.2 lakh
Practical annual revenue opportunityRs 120 crore

This may be less glamorous than a report-based TAM, but it tells you what kind of company you may actually build.

Use a worksheet instead of one heroic TAM number.

InputConservativeBaseOptimistic
Reachable customers in beachhead
Percentage reachable through chosen channels
Adoption over 3-5 years
Annual revenue per customer
Gross margin
Sales and onboarding cost
Churn or replacement rate
Expansion revenue

Then calculate:

reachable customers x adoption x annual revenue x gross margin

This is less exciting than a giant market report, but it is much harder to fool yourself. If the business only works in the optimistic case, the founder needs stronger evidence before hiring or raising against the plan.

For the first market, build a real list before trusting the model.

Ask:

  • Can we name 100 companies or users in this segment?
  • Can we find contact paths for at least 50?
  • Can we get 20 qualified conversations without paid fantasy channels?
  • Do at least 10 show the same pain?
  • Do at least 5 show commitment through time, data, payment, or pilot interest?

If you cannot build a prospect list, the market may still be real, but it is not yet reachable by you. Reachability is part of market size. A customer you cannot find is not part of your obtainable market today.

For B2B, market size often comes from existing budgets. Ask:

  • What budget line does this replace or create?
  • Is the buyer already spending on software, people, agencies, tools, compliance, ads, support, or manual operations?
  • What is the cost of the current workaround?
  • Who owns that budget?
  • Is the spend recurring or one-time?

If the problem has no budget, you may need to sell strategic value, risk reduction, revenue lift, or cost savings. That is possible, but harder than shifting existing spend.

Use more than one model so one fantasy assumption does not control the answer.

Number of target customers x annual revenue per customer

Useful for SaaS, services, and B2B tools.

Number of transactions x take rate or fee

Useful for marketplaces, fintech, logistics, commerce, and payment-linked products.

Number of workflows or tasks x value per workflow

Useful for AI, automation, compliance, back-office, support, and operational tools.

If all three models give wildly different answers, that is not a problem. It is a signal to understand the business model more deeply.

Most early sizing models are wrong. The question is which assumption matters most.

Stress-test:

AssumptionWhat if it is half as good?
Number of reachable customersDoes the beachhead still justify focus?
PriceCan the business support sales and support?
AdoptionIs the channel or urgency weaker than expected?
Gross marginIs delivery too service-heavy?
ChurnDoes acquisition need to be much cheaper?
Sales cycleCan runway survive the buying process?
Support loadDoes price match customer effort?

Founders often discover that the market is not too small; the model is too fragile. For example, a SaaS plan may look strong at Rs 1 lakh ACV and 20% adoption, but weak at Rs 36,000 ACV, 8% adoption, and high support. That does not mean the startup is bad. It means pricing, segment, onboarding, or GTM must change.

For consumer products, user count is not enough. You need monetizable behavior.

Estimate:

  • Target users.
  • Reachable users.
  • Activation rate.
  • Retention rate.
  • Paying conversion or revenue per active user.
  • Acquisition cost.
  • Gross margin.

Large user numbers can hide tiny revenue if willingness to pay is weak, ad inventory is low value, or retention is poor.

A market that supports Rs 100 crore of high-margin recurring software revenue is different from Rs 100 crore of low-margin, service-heavy, support-heavy revenue. Include gross margin and operating burden in your thinking.

Ask:

  • Does revenue repeat?
  • Does usage expand?
  • Does servicing cost rise with revenue?
  • Does the customer segment churn?
  • Does collection take too long?
  • Does the product require heavy customization?

Your first market is not your final market. Strong startups often begin narrow and expand.

Think in layers:

  1. First customer type.
  2. First workflow.
  3. First geography.
  4. Adjacent customer type.
  5. Adjacent workflow.
  6. Larger platform or product line.

Example:

  • Start: compliance workflow for funded Indian startups.
  • Expand: compliance and finance operations for Indian SMEs.
  • Expand further: finance automation for companies selling globally from India.

The first wedge must be specific. The future story must be plausible.

A good expansion story has adjacency, not fantasy.

Strong expansion:

  • Same buyer, adjacent workflow.
  • Same workflow, adjacent customer segment.
  • Same customer, higher-value module.
  • Same category, new geography with similar behavior.
  • Same data asset, new insight product.

Weak expansion:

  • “Then we sell to everyone.”
  • “Then we launch in all countries.”
  • “Then we add AI.”
  • “Then large enterprises will buy.”
  • “Then we become a platform.”

Investors and founders should both ask: what proof from the first market gives us the right to enter the second?

India has large informal markets. Spend may exist but not appear in clean data. Customers may pay agents, staff, brokers, cash vendors, or family networks instead of software.

Tier 1 behavior cannot represent India. Pricing, language, trust, payment behavior, and support expectations can change outside metros.

Some segments are smartphone-heavy but software-light. Others use digital tools but still rely on manual reconciliation, WhatsApp, Excel, or phone calls.

A large number of users does not mean a large revenue market. Ask what customers already pay for and what budget line you can replace.

If customers require field sales, implementation, training, or local trust, your obtainable market may be much smaller than your theoretical market.

Public data may be outdated, inconsistent, or too broad. Combine reports with primary interviews and observable business lists.

Some Indian markets have many users but few real buyers. Students may use a product, but parents, colleges, employers, or coaching institutes may pay. Users can be large while budget is concentrated elsewhere.

Revenue potential is not the same as cash collected. Long payment cycles, TDS, GST issues, purchase order processes, and informal negotiations can reduce practical market size.

High-touch onboarding, WhatsApp support, training, language needs, and field operations can make a large market less attractive if price points are low.

  1. Define the beachhead segment.
  2. Estimate how many such customers exist.
  3. Estimate how many are reachable through your channels.
  4. Estimate realistic annual revenue per customer.
  5. Estimate adoption over 3-5 years.
  6. Compare gross margin and sales cost.
  7. List expansion segments.
  8. State assumptions clearly.

After building the model, stress-test it:

  • Can we name 50 real customers in the first segment?
  • Can we reach them through a channel we control or can afford?
  • Does the assumed price match current spend or visible pain?
  • Does gross margin support sales, onboarding, support, and collections?
  • Does adoption require behavior change that the model ignores?
  • Does the first market produce enough proof for the next market?
  • Is the expansion path adjacent or imaginary?
  • Are we counting users when the payer is someone else?
  • Are we assuming US or global ARPU in an Indian segment without evidence?

You may eventually show a clean TAM/SAM/SOM slide. Internally, keep a messier founder model.

Investor slideFounder model
Shows the large opportunity.Shows the first reachable market.
Uses simple categories.Uses segment, channel, price, margin, and adoption assumptions.
Explains expansion story.Tests whether expansion is adjacent and earned.
Communicates ambition.Guides hiring, burn, pricing, and GTM.

Do not let the investor slide become the operating plan. The slide can show the mountain. The founder model must show the first climb.

A smaller, honest model is better than a large model that hides weak assumptions.

Not every good market is a venture-scale market. That is fine, but founders must know the difference.

Venture-scale usually needs:

  • Large reachable market.
  • Fast growth potential.
  • High gross margins or strong marketplace economics.
  • Repeatable acquisition.
  • Retention or network effects.
  • Ability to deploy capital efficiently.
  • Expansion beyond the first niche.

A good non-venture business may have:

  • Smaller market.
  • Strong cash flow.
  • Services-heavy delivery.
  • Slower but profitable growth.
  • Founder-led sales.
  • High customer trust and low churn.

The danger is raising venture money for a market that wants a slower, profitable company. Market sizing should inform financing strategy.

Label every number:

  • Evidence: based on customer list, invoices, usage, interviews, paid pilots, government data, or reliable reports.
  • Estimate: based on partial evidence and reasonable extrapolation.
  • Guess: plausible but unverified.

Your next discovery work should replace guesses with evidence. A sizing model that admits uncertainty is more useful than a confident spreadsheet built on fiction.

Suppose you are building software for small diagnostic labs in India.

Start narrow:

  • Segment: independent labs in three metro regions.
  • Number of likely reachable labs: 2,000.
  • Early adoption over five years: 15%.
  • Annual revenue per lab: Rs 60,000.
  • Revenue opportunity: Rs 1.8 crore annually in the first wedge.

That may be too small for a venture-scale company by itself. Then test expansion: multi-branch labs, clinics with in-house diagnostics, billing and reporting modules, inventory workflows, patient communication, and integrations if relevant.

Now the founder can see whether the first wedge is a learning wedge, a cash-flow business, or the start of a larger platform.

The most useful part of market sizing is not the final number. It is seeing which assumptions control the number. A founder should know which assumptions are solid, which are fragile, and which need fieldwork.

Create an assumption table:

AssumptionCurrent NumberEvidence LevelRisk If WrongNext Validation Action
Reachable customers in wedgeEvidence / estimate / guessBuild named prospect list.
Decision makers who feel urgencyEvidence / estimate / guessInterview buyers, not only users.
Annual price customer will acceptEvidence / estimate / guessTest paid pilot or proposal.
Sales cycleEvidence / estimate / guessTrack first 10 real deals.
Gross marginEvidence / estimate / guessModel onboarding, support, infrastructure, and service cost.
RetentionEvidence / estimate / guessTrack usage after first value.
Expansion pathEvidence / estimate / guessIdentify adjacent workflows or segments.

Do not hide weak assumptions in a beautiful spreadsheet. Highlight them. Your next operating plan should attack the weakest high-impact assumptions first.

The first market does not need to be the whole company, but it must be honest about its job.

There are four common jobs for the first wedge:

Wedge JobWhat It MeansFinancing Implication
Learning wedgeSmall but reachable segment that teaches the product and buyer.Keep burn low; do not over-hire for scale.
Cash-flow wedgeSegment can produce profitable early revenue.Useful for bootstrapping or extending runway.
Reference wedgeSegment creates logos, case studies, or credibility for a bigger adjacent market.Measure transferability of references.
Platform wedgeSegment creates data, supply, demand, integrations, or network density.Needs proof that the wedge compounds.

A weak first wedge is one that is small, hard to reach, low-margin, non-referential, and does not lead anywhere. A strong first wedge may be small but gives the company leverage: learning, revenue, trust, data, distribution, or expansion.

After sizing the first wedge, write the expansion chain:

  1. First exact segment.
  2. Adjacent customer with similar workflow.
  3. Adjacent workflow inside the same customer.
  4. Adjacent geography or language market.
  5. Adjacent product line or monetization layer.

For each step, write what must be true before expansion:

Expansion StepWhat Must Be TrueProof Needed
Segment 1 to segment 2Pain and buying process are similar.References transfer and sales message still works.
Workflow 1 to workflow 2Customer trusts you inside the operation.Customers request the adjacent workflow and will pay.
City or region expansionChannel and support can repeat.Acquisition, onboarding, and retention stay healthy.
Product-line expansionThe first product has retention and credibility.Existing customers adopt the second product.

This protects you from fake TAM thinking. Expansion is not a slide. Expansion is a sequence of earned permissions.

A good market-size model should be easy to attack. If the model only confirms your optimism, it is not yet a founder tool.

List the assumptions that, if wrong, would break the opportunity:

AssumptionOptimistic VersionConservative VersionWhat Would Prove It Wrong?Test
Number of reachable customersBuild actual prospect list.
Buyer existenceInterview budget owners.
PriceAsk for paid pilot or invoice-level commitment.
Adoption rateTrack conversion from qualified conversations.
RetentionFollow usage after first value.
Sales costMeasure founder hours and channel cost per qualified buyer.
Support loadTrack onboarding, training, and exception handling.
Expansion pathAsk whether adjacent segment accepts current proof.

Then choose the three assumptions with the highest impact and weakest evidence. Those are not spreadsheet cells. They are next week’s research agenda.

Ask:

  • If the market is as large as we claim, why has the problem not already been solved well?
  • What changed that makes the opportunity newly accessible?
  • Which customers can we actually reach in the next 90 days?
  • What price can we defend without fantasy ROI?
  • Which segment is big enough for the next funding or bootstrapping milestone?
  • What would make this a good business but a poor venture-backed business?
  • What hidden service, support, compliance, or distribution cost shrinks the real market?

Investors may ask for TAM. Founders need something more useful: a believable sequence from first wedge to durable company.

The most practical market-size number in the early days is not TAM. It is the number of customers you can credibly identify, contact, qualify, and serve.

Build a reachable market audit:

LayerCountEvidence SourceConfidence
Total theoretical customersReports, registries, platforms, public data.Low / medium / high
Customers matching your narrow segmentFilters, directories, associations, communities.Low / medium / high
Customers you can nameProspect list.Low / medium / high
Customers you can reachEmail, phone, intro, community, partner, event.Low / medium / high
Customers with urgent triggerRecent hiring, compliance, growth, complaint, migration, budget.Low / medium / high
Customers likely to pay nowDiscovery, pilots, past spend, purchase intent.Low / medium / high

If a market has 100,000 theoretical customers but only 50 you can name and 5 you can reach, the immediate market is not 100,000. Your first job is distribution learning, not scale hiring.

Indian market data can be noisy. Categories may be informal, public datasets may lag, and businesses may not describe themselves the way software founders segment them. Use multiple weak signals rather than pretending one report is precise:

  • Public registries and directories.
  • Trade associations.
  • GST or compliance maturity signals where appropriate and lawful.
  • Job postings.
  • Marketplace listings.
  • Local consultants and service providers.
  • WhatsApp, community, and regional networks.
  • Payment behavior and invoice reality from discovery.

Do not turn noisy data into fake precision. Mark confidence openly.

Early founders often count customers equally. That hides the business reality. A customer who is easy to reach, pays faster, needs little support, and renews is not equal to a customer who looks large but takes months of founder time.

Build a revenue-weighted sizing view:

SegmentTheoretical CountReachable CountLikely PriceSales EffortSupport LoadGross Margin RiskConfidence
Low / medium / highLow / medium / highLow / medium / highEvidence / estimate / guess

Then add a founder-hours view:

Expected first-year revenue per customer
- expected acquisition effort
- onboarding/support effort
- discounting/payment delay risk
= practical attractiveness of the segment

This is not a finance model. It is a reality check. In many Indian markets, the segment with the biggest logo value may not be the best first market because sales cycles, payment terms, implementation, and support can consume the company.

Use revenue-weighted sizing when choosing between:

  • Enterprise versus SMB.
  • Tier 1 versus Tier 2 or smaller-city customers.
  • Software-only versus service-assisted delivery.
  • Direct sales versus partner-led distribution.
  • India-first versus global-from-India wedge.

A small reachable segment with high urgency, fast sales, repeatable onboarding, and references can be a better first market than a giant theoretical market with slow trust and unclear payment.

Market sizing should control how aggressively the company spends. A founder should not hire, raise, or expand as if the future market is already proven when only the first wedge is understood.

Use a sizing-to-burn guardrail:

Market evidenceWhat it supportsWhat it does not yet support
Report-level TAM onlyDiscovery, interviews, small experiments.Hiring a large team or claiming repeatable GTM.
Named reachable accountsFounder-led sales, narrow product wedge, low-burn validation.Paid scale or broad market expansion.
Pain-qualified buyersMVP, paid pilots, sharper positioning.Heavy sales team or multi-city rollout.
Paid pilots or early contractsSmall customer success motion, onboarding investment.Assuming retention, expansion, or venture-scale economics.
Retained customers with repeatable acquisitionHiring around a known motion.Expanding into unrelated segments.
Segment expansion proofMore aggressive fundraising and GTM.Treating every adjacent market as automatically unlocked.

The founder’s budget should match the evidence level. If market evidence is low-confidence, keep burn low and learn. If market evidence is strong, spend to exploit what is working.

Before hiring for growth, ask:

HireMarket-size proof needed
SDR or outbound teamNamed reachable accounts, clear ICP, working message, repeatable response signal.
Sales leaderFounder has closed enough similar deals to show process, objections, cycle, and pricing.
Customer successOnboarding and retention require repeatable human effort, not random custom rescue.
Paid marketingChannel can produce qualified buyers, not only cheap leads.
Implementation teamImplementation is part of the model and priced into gross margin.
Expansion or partnerships roleFirst wedge references transfer to adjacent segment or channel.

If the answer is “we need the hire to discover the market,” be careful. That may be valid for a short experiment, but it is not the same as hiring into a proven market.

When fundraising, separate three claims:

ClaimEvidence required
The market can be largeTop-down context, adjacent budgets, structural change, expansion logic.
We can enter itNamed accounts, interviews, pilots, first sales, founder access, channel signal.
We can scale inside itRepeatable acquisition, retention, gross margin, sales cycle, references, expansion proof.

A deck can show all three. But the founder should know which claim is proven and which is still a hypothesis. This honesty improves the pitch because it shows command of the business, not lack of ambition.

In India-first markets, a large theoretical market can tempt founders into premature field sales, support hiring, local language expansion, city launches, channel partnerships, or heavy implementation capacity. These costs can arrive before the market has proven willingness to pay.

Use this rule:

Do not build a cost structure for the TAM. Build a cost structure for the evidence-backed wedge.

Then expand the cost structure only when the next segment has earned proof: reachable buyers, paid demand, retention, references, and manageable support economics.

  • Using huge top-down numbers without a wedge.
  • Claiming “1% of a huge market” as strategy.
  • Ignoring willingness to pay.
  • Treating users as buyers.
  • Forgetting sales and support cost.
  • Combining unrelated segments into one number.
  • Making a VC-scale claim for a non-VC market.
  • Ignoring gross margin.
  • Forgetting that reachable customers are fewer than theoretical customers.
  • Treating a service-heavy opportunity as software-like revenue.
  • Assuming global ARPU applies to Indian segments.
  • Hiding weak first-market size behind a giant future market.

Build a one-page sizing model:

SegmentCustomersReachable %Price/YearAdoption %Revenue Potential

Fill it with your best current assumptions. Mark each assumption as evidence, estimate, or guess. Your next research task is to replace guesses with evidence.

After sizing the market, filter it down to demand you can actually reach, sell to, serve, and learn from. A large market is not useful if the first reachable customers are not ready to change.

Use this filter:

AssumptionCurrent valueEvidence qualityHow to test
Number of target customersEvidence / estimate / guessDirectory, industry data, scraped list, associations, manual count.
Reachable percentageEvidence / estimate / guessOutreach tests, channel tests, referral path, search volume.
Buyer budgetEvidence / estimate / guessDiscovery, current spend, invoices, procurement conversations.
Willingness to payEvidence / estimate / guessPaid pilots, price tests, proposal reactions.
Adoption rateEvidence / estimate / guessConversion from outreach, pilot-to-paid, onboarding completion.
RetentionEvidence / estimate / guessRepeat usage, renewal intent, churn calls, cohort data.
Gross marginEvidence / estimate / guessDelivery effort, support load, infra cost, services component.
Sales cycleEvidence / estimate / guessDays from first conversation to payment or signed pilot.

The most important column is not the number. It is evidence quality. A founder should know whether the company is acting on a real market, a reasonable estimate, or a story.

Separate the first market from the future market:

MarketPurposeQuestion
BeachheadWhere you get proof.Can this segment produce real customer evidence within months?
ExpansionWhere proof can travel.Which adjacent segment trusts the first segment’s proof?
Long-term marketWhere the company could become large.What must become true before this larger claim matters?

Investors may care about the large market. Founders must care about the first market. You cannot compound into the future market if the beachhead does not buy, use, retain, or refer.

For India, do not count need as demand. Many people may need a solution but still not have budget, trust, workflow readiness, payment behavior, or decision authority. Mark the gap between:

  • People/businesses with the problem.
  • People/businesses aware of the problem.
  • People/businesses willing to change.
  • People/businesses reachable by you.
  • People/businesses able and willing to pay now.

The last number is usually much smaller, and much more useful.

Treat market sizing as an evidence ladder, not a single spreadsheet.

LevelEvidenceFounder confidence
1. Theoretical countReports, directories, public databases, broad industry estimates.Useful for orientation, weak for operating decisions.
2. Reachable countNamed accounts, communities, associations, search results, channel lists.Useful for first GTM planning.
3. Pain-qualified countCustomers who show recent pain and current workaround.Useful for narrowing segment.
4. Buyer-qualified countPain plus budget owner, approval path, and trust requirement.Useful for sales planning.
5. Commercially qualified countPricing, pilot, proposal, paid test, or procurement signal.Useful for revenue forecast.
6. Retained countCustomers who continue using and paying.Useful for real market quality.

Most early founders stop at level 1. Useful founders climb toward levels 3-5 before making big product, hiring, or fundraising claims.

Build the first market from the bottom up.

Exact segment:
Named reachable accounts:
Expected response rate:
Discovery-qualified accounts:
Pilot-qualified accounts:
Expected close rate:
Average first-year revenue:
Expected gross margin:
Expected sales/support effort:

Then translate to a simple model:

InputValueEvidence quality
Named accounts reachable in 90 daysEvidence / estimate / guess
Useful conversation rateEvidence / estimate / guess
Pilot/proposal rateEvidence / estimate / guess
Close rateEvidence / estimate / guess
Average first-year revenueEvidence / estimate / guess
Support/load riskEvidence / estimate / guess

Example:

We can identify 600 export manufacturers in two clusters, reach 200 through associations and founder outreach, speak to 40, run 8 paid diagnostics, close 3 annual customers at Rs 3-5 lakh each, and use those references to expand into adjacent clusters.

This is a better first-market claim than “India has millions of SMEs.”

A market-size slide should connect the big market to the first wedge.

Use three layers:

LayerWhat to show
Big marketWhy the problem can become large if the company wins.
Serviceable marketWhich segments have the budget, workflow, and adoption readiness.
BeachheadWhich first customers are reachable now and why they will produce proof.

The slide should answer:

  • Why is the opportunity large enough eventually?
  • Why is the first segment narrow enough to win?
  • What proof from the first segment can travel to the next segment?
  • Which assumptions are still unproven?

Do not hide a weak beachhead behind a huge TAM. A serious investor or advisor will ask how the founder gets from first customer to large market.

Not every market-size number deserves the same confidence. Label your numbers so the team does not treat guesses like facts.

ConfidenceSourceHow to use
LowReports, broad assumptions, social media, rough estimates.Directional story only. Do not hire or spend heavily from this.
MediumNamed prospect list, customer interviews, budget clues, channel data.Useful for beachhead planning and experiment design.
HighPaid pilots, conversion rates, retained customers, sales cycle, gross margin.Useful for operating plan and fundraising narrative.

For every sizing model, write:

The number we believe:
Source:
Confidence:
Largest assumption:
What would make this number smaller:
What evidence will improve confidence:
Decision this number supports:

The last line is important. A market-size number should support a decision: enter, narrow, fundraise, hire, build, expand, or stop. If it supports no decision, it may be decoration.

TAM can be huge while the first serviceable market is small. Founders need both.

Define the serviceable first market:

FilterQuestion
ReachCan we identify and contact these customers within 90 days?
PainDo they have the problem now, not someday?
BudgetCan someone approve payment?
TrustCan we create enough credibility to start?
WorkflowCan our product or service deliver value without heavy custom work?
SupportCan we support them at current team size?
ReferenceWill one win help sell the next similar customer?

If the first serviceable market cannot produce enough learning, revenue, or proof, the company may need a different wedge even if the long-term market is large.

Do not say:

The market is 10,000 crore, so getting 1 percent is enough.

Say:

There are 800 reachable accounts in our first wedge. We can reach 200 this quarter, expect 40 serious conversations, 8 pilots, and 3-5 paying customers if the pain and trust assumptions hold.

That second statement is smaller, but it can be tested.