Market Sizing Questions: How to Solve Any Question (With Examples)

Your interviewer says: "How many electric vehicle charging stations will the US need by 2030?"

You have 60 seconds before the silence gets uncomfortable. No Google. No spreadsheets. Just your brain, a pen, and whatever structure you can pull together on the spot.

Market sizing questions are the part of consulting interviews that terrify quantitative thinkers and non-quant candidates alike — but for completely different reasons. The math people overcomplicate it. Everyone else freezes because they don't know where to start.

Here's what both groups miss: market sizing isn't a math test. It's a structured thinking test that happens to involve numbers. The interviewer doesn't care whether your estimate is $4.2 billion or $6.8 billion. They care whether your approach to getting there was logical, transparent, and defensible.

This guide gives you the universal approach to market sizing, five fully worked examples covering different question types, and a cheat sheet you can reference the night before your interview. No filler. No "first, let's define what market sizing is" padding.

[INTERNAL LINK: case interview frameworks]

Key Takeaways (TL;DR)

• Every market sizing question follows the same pattern: break a big unknown into smaller knowable pieces, estimate each piece, multiply/add them together, and sanity-check the result
• Top-down vs. bottom-up isn't a choice — use whichever the question lends itself to, or use both as a cross-check
• Your assumptions matter more than your math. State them explicitly. An interviewer who disagrees with your assumption will redirect you — that's a feature, not a failure
• Precision is a trap. Saying "approximately 7%" is better than saying "7.3%" when you're estimating. False precision signals that you don't understand the exercise
• Always sanity-check. Candidates who skip the final "does this make sense?" step leave points on the table in over 60% of cases
• Practice 10-15 market sizing questions and you'll have seen every structural pattern that exists. After that, returns diminish rapidly

The Universal Approach to Market Sizing

Every market sizing question — whether it's "how many golf balls fit in a school bus" or "what's the total addressable market for cloud kitchen software in Southeast Asia" — follows the same four-step process:

Step 1: Clarify and Scope

Before you touch a single number, make sure you know what you're solving for.

"How big is the coffee market?" is actually three different questions:

• Revenue? Volume (cups)? Number of establishments?
• US only? Global? Urban areas?
• Just coffee shops? Including grocery retail? At-home consumption?

Spending 15 seconds asking one or two clarifying questions is never wrong. It signals maturity and prevents you from solving the wrong problem — which happens to roughly 20% of candidates in market sizing questions according to interviewer feedback data.

Step 2: Choose Your Approach and Build the Tree

You have two tools: top-down and bottom-up. Your job is to build an estimation tree — a logical decomposition that breaks your big unknown into smaller, estimable pieces.

Top-down: Start with a large known number and apply successive filters.

Total population → relevant segment → usage rate → frequency → price = market size

Bottom-up: Start with a single unit and scale up.

One store × transactions/day × average ticket × days/year × number of stores = market size

Neither approach is "better." The right choice depends on the question. Sometimes you'll use both as a cross-check — and when they converge within 20-30% of each other, you can be confident in your estimate.

Step 3: Estimate Each Component

This is where most candidates get nervous, but it's actually the easiest part if you've built a good tree. Each component should be something you can estimate with reasonable confidence:

• Known quantities: US population (330 million), number of US households (130 million), average household size (~2.5), days in a year (365)
• Common sense estimates: % of people who own cars (~90% of US adults), average restaurant meal price ($15-25), hours in a workday (8)
• Educated guesses: % of households with a piano (5%), frequency of dental visits (2x/year), % of businesses using cloud software (60%)

The golden rule: State every assumption out loud. "I'm going to assume about 5% of households own a piano — does that feel reasonable?" This does two things: it makes your logic transparent, and it gives the interviewer a chance to course-correct you. Research on case interview scoring shows that explicit assumption-stating correlates with 30% higher structure scores compared to candidates who embed assumptions silently.

Step 4: Calculate and Sanity-Check

Do the math — keeping it clean and rounded. Then ask yourself: does this answer make sense?

Sanity checks include:

• Per capita reasonableness: If your US market size divided by 330 million people gives an implausible per-person spend, something's off
• Comparison to known markets: If you estimated the US piano tuning market at $50 billion, that's larger than the music instrument industry — red flag
• Order of magnitude: Is your number in the right ballpark of millions, billions, or trillions?

About 40% of candidates skip the sanity check entirely. Don't be one of them. A quick "That gives us roughly $X — let me check if that's reasonable by comparing it to Y" takes 10 seconds and significantly improves your interviewer's assessment.

[INTERNAL LINK: case interview tips]

Top-Down vs. Bottom-Up: When to Use Each

This distinction matters less than most prep resources suggest, but here's when each shines:

Top-Down Works Best When:

• You're sizing a consumer market where population is a natural starting point
• The question involves penetration rates or adoption curves
• You need a quick, defensible estimate and don't have granular unit economics
• Example: "What's the market size for contact lenses in the US?"

Bottom-Up Works Best When:

• You're sizing a B2B market with identifiable customer segments
• The question involves physical capacity or operational constraints
• You have a good mental model of the unit economics of a single entity
• Example: "What's the annual revenue of all McDonald's locations in the US?"

Use Both When:

• The interviewer seems to want a thorough answer
• You have time and want to cross-validate
• Your initial estimate feels uncertain and you want a second angle

Studies of high-performing candidates show that those who cross-validate with both approaches score approximately 25% higher on quantitative rigor than those who rely on a single method. It's a simple differentiator that most candidates miss.

Five Worked Examples (With Full Math)

Example 1: Consumer Product (Top-Down)

Question: "How many disposable diapers are sold in the US each year?"

Approach: Top-down from population.

Step 1: How many babies/toddlers are in diapers?
  - US births per year: ~3.6 million
  - Children in diapers: ages 0–2.5 (approximately)
  - So: 3.6M × 2.5 years = ~9 million children in diapers at any given time

Step 2: How many diapers does one child use per day?
  - Newborns: ~10/day
  - 6-12 months: ~8/day
  - 1-2 years: ~6/day
  - Weighted average across the age range: ~7 diapers/day

Step 3: Annual consumption per child
  - 7 diapers/day × 365 days = ~2,555 diapers/year
  - Round to ~2,500

Step 4: Total market volume
  - 9 million children × 2,500 diapers = 22.5 billion diapers/year

Step 5: Market size in revenue
  - Average price per diaper: ~$0.25 (bulk pricing)
  - 22.5 billion × $0.25 = ~$5.6 billion

Sanity check: The US diaper market is widely reported at $5-7 billion.
Our estimate of $5.6 billion falls right in that range. ✓

What makes this strong: Clean tree, explicit assumptions at each step, and a sanity check that confirms we're in the right ballpark. This is exactly the kind of answer that scores well.

Example 2: B2B Market (Bottom-Up)

Question: "What's the market size for commercial coffee machines in US offices?"

Approach: Bottom-up from the customer base.

Step 1: How many offices are there in the US?
  - US workforce: ~160 million
  - Roughly 60% work in office settings: ~96 million office workers
  - Average office size: ~50 employees
  - Number of offices: 96M ÷ 50 = ~1.9 million offices
  - Round to ~2 million offices

Step 2: What % of offices have a commercial coffee machine?
  - Small offices (<20 people): maybe 30% have commercial machines (rest use consumer)
  - Medium offices (20-100): ~70% have commercial machines
  - Large offices (100+): ~90% have commercial machines
  - Weighted estimate (skewing toward more small offices): ~50%
  - Offices with commercial machines: 2M × 50% = 1 million

Step 3: How many machines per office?
  - Small offices: 1 machine
  - Large offices: 2-3 machines (one per floor/wing)
  - Average: ~1.3 machines per office
  - Total installed base: 1M × 1.3 = ~1.3 million machines

Step 4: Annual market (replacement + new)
  - Machine lifespan: ~7 years
  - Annual replacement: 1.3M ÷ 7 = ~185,000 machines
  - New office growth: ~2% per year = ~26,000 new machines
  - Total annual sales: ~210,000 machines

Step 5: Revenue
  - Average commercial coffee machine price: ~$3,000
  - 210,000 × $3,000 = ~$630 million (machines only)
  - Add service contracts + supplies (typically 2x the machine revenue in B2B): 
    ~$1.3 billion total market including recurring revenue

Sanity check: This suggests each office worker's share is about 
$1,300M ÷ 96M = ~$14/year on office coffee equipment and supplies. 
That feels low but reasonable given most of the cost is amortized. ✓

What makes this strong: It demonstrates bottom-up thinking for a B2B market where starting with population wouldn't make sense. The distinction between installed base and annual sales is sophisticated — most candidates forget that market sizing for durable goods needs to account for replacement cycles.

Example 3: Geographic Market

Question: "How many Starbucks locations could profitably operate in the city of Chicago?"

Approach: Demand-side analysis with geographic density.

Step 1: Total demand for premium coffee in Chicago
  - Chicago population: ~2.7 million (city proper)
  - Add commuters/visitors: effective daytime population ~3.2 million
  - % who are premium coffee drinkers: ~35%
  - Premium coffee drinkers in Chicago: 3.2M × 35% = ~1.12 million

Step 2: Frequency and spend
  - Average premium coffee visits per week: ~3
  - Annual visits per person: 3 × 52 = 156
  - Average ticket: ~$5.50
  - Total premium coffee market: 1.12M × 156 × $5.50 = ~$960 million/year

Step 3: Starbucks' share of premium coffee
  - Starbucks typically captures 35-40% of the premium coffee segment in major US cities
  - Starbucks revenue in Chicago: $960M × 37% = ~$355 million

Step 4: Revenue needed per profitable location
  - Average profitable Starbucks: ~$1.5-2 million annual revenue
  - Using $1.7M as a midpoint

Step 5: Number of profitable locations
  - $355M ÷ $1.7M = ~209 locations

Sanity check: Chicago currently has approximately 200-300 Starbucks locations
depending on how you count (company-operated vs. licensed). 
Our estimate of ~209 company-operated profitable locations tracks well. ✓

What makes this strong: It starts with the demand side rather than supply, naturally incorporates competition (Starbucks' share), and uses a profitability threshold — not just "how many could physically fit" but "how many could make money." That's a more sophisticated question to answer.

Example 4: TAM/SAM/SOM

Question: "What's the TAM, SAM, and SOM for a new AI-powered legal document review platform targeting US mid-size law firms?"

Approach: Nested market scoping — broad to narrow.

TAM (Total Addressable Market): All legal document review spending globally

  - Global legal services market: ~$1 trillion
  - Document review as % of legal work: ~15-20% (discovery, due diligence, compliance)
  - Global legal document review: $1T × 17% = ~$170 billion
  
  TAM = ~$170 billion

SAM (Serviceable Addressable Market): AI-enabled document review for US law firms

  - US share of global legal market: ~40%
  - US legal document review: $170B × 40% = ~$68 billion
  - % of document review addressable by AI tools (not all work can be automated): ~40%
  - AI-addressable document review in US: $68B × 40% = ~$27 billion
  - Further filter: law firms only (exclude corporate in-house): ~60% of market
  
  SAM = ~$16 billion

SOM (Serviceable Obtainable Market): What this startup can realistically capture

  - Target: mid-size US law firms (50-300 attorneys)
  - Mid-size firms as % of total law firm revenue: ~25%
  - Mid-size law firm AI document review: $16B × 25% = ~$4 billion
  - Realistic market share for a new entrant in Year 3: 1-3%
  
  SOM (Year 3) = ~$40-120 million
  Using midpoint: SOM = ~$80 million

Sanity check: 
- TAM of $170B for global legal document review? Legal is a $1T industry 
  and document review is a massive component — plausible. ✓
- SOM of $80M implies roughly 50-100 mid-size firm clients at $800K-1.6M 
  average contract value. For an AI platform with real ROI, 
  that's achievable but ambitious. ✓

What makes this strong: It cleanly distinguishes TAM/SAM/SOM — which most candidates conflate — and applies defensible filters at each level. The SOM calculation is particularly thoughtful because it includes a realistic market share assumption rather than claiming "we can capture 10% of a $170B market" (a classic startup pitch error that interviewers will skewer).

[INTERNAL LINK: case interview examples]

Example 5: "How Many" Estimation

Question: "How many weddings happen in the US each year?"

Approach: Top-down demographic analysis.

Step 1: Marriage-eligible population
  - US population: ~330 million
  - Adults (18+): ~260 million
  - Currently unmarried adults: ~45% → ~117 million
  
Step 2: Annual marriage rate
  - Each marriage requires two people
  - Marriage rate: historically ~6-7 marriages per 1,000 population
  - 330M × 6.5/1,000 = ~2.15 million marriages per year
  - Each marriage = one wedding (some elope, but some have multiple ceremonies)
  
Step 3: Adjust for modern trends
  - Marriage rates have declined ~15% over the past decade
  - But post-COVID saw a wedding boom with backlogs
  - In 2025-2026, we're likely at a normalized rate
  - Estimate: ~2.0-2.2 million weddings per year

Final estimate: ~2.1 million weddings per year

Sanity check: The wedding industry frequently cites 2.0-2.5 million 
weddings annually in the US. Our estimate of 2.1 million is right 
in the middle. ✓

Bonus — market sizing the wedding industry:
  - 2.1 million weddings × ~$35,000 average wedding cost 
  - = ~$73.5 billion US wedding industry
  - (Industry reports cite $70-80 billion — confirmed ✓)

What makes this strong: It's clean and fast — this is a 2-minute answer, which is appropriate for a simpler estimation question. The bonus extension to market size shows initiative without being asked. Notice the estimate uses the demographic approach (marriage rate per 1,000 population) rather than trying to guess what percentage of singles get married each year — a subtle choice that produces a more reliable estimate.

The Market Sizing Cheat Sheet

Keep these reference numbers in your head. You don't need to memorize them precisely — rough ranges are fine:

US Demographics

Business & Economic

Useful Conversion Factors

Quick Mental Math Tips

• 10% of 330M = 33 million (useful for any "what % of Americans" question)
• Doubling rule: If something grows at X% per year, it doubles in roughly 70/X years
• Round aggressively: Use 300M instead of 330M if it makes the math cleaner. Nobody will penalize you.
• Work in millions, not raw numbers: "3.6 million births × 2,500 diapers = 9 billion diapers" is cleaner than trying to multiply 3,600,000 × 2,500

The Five Most Common Market Sizing Traps

Trap 1: Solving for the Wrong Thing

"How big is the coffee market?" could mean revenue, volume, number of shops, or total cups consumed. If you don't clarify, you might produce a perfectly logical answer to the wrong question. About 1 in 5 candidates makes this mistake — and it's the most preventable one.

Trap 2: False Precision

Saying "I estimate 7.3% of the population..." when you're pulling a number from thin air. Round numbers signal that you understand you're estimating. Precise numbers signal that you think you're calculating. In an estimation exercise, the first is a strength and the second is a red flag.

Trap 3: Forgetting to Segment

"The average American eats X pizzas per year" ignores that children eat fewer than adults, coastal cities have different eating habits than rural areas, and frequency varies enormously by age group. Smart segmentation — even just splitting into 2-3 groups — dramatically improves accuracy and shows structured thinking.

Trap 4: Missing the Replacement Cycle

For durable goods (cars, appliances, commercial equipment), the annual market isn't the total installed base — it's new units + replacements. A $500 million installed base of machines with a 10-year lifespan means the annual market is roughly $50 million in replacements plus whatever growth adds. Candidates who forget this overestimate annual markets by 5-10x.

Trap 5: No Sanity Check

You calculate that the US market for dog grooming is $500 billion. That's larger than the US restaurant industry. Something went wrong — but if you don't sanity-check, you'll present this number with confidence and your interviewer will wonder if you have any business intuition at all.

[INTERNAL LINK: how to practice case interviews]

How to Practice Market Sizing (Efficiently)

You don't need to do 100 market sizing questions. That's diminishing returns territory. Here's the efficient path:

1. Do 5 questions slowly — take 10-15 minutes each, write out full trees, check your answers against industry data
2. Do 5 questions at interview speed — 3-5 minutes, speaking out loud, simulating the real pressure
3. Do 5 questions with feedback — either with a partner or using a tool that can evaluate your structure and assumptions

After those 15 questions, you'll have seen every structural pattern: consumer products, B2B services, geographic markets, "how many" estimations, and TAM/SAM/SOM breakdowns. Additional practice has marginal value unless you're specifically weak on mental math. Tools like Kasie and PrepLounge can provide the realistic practice environment and calibrated feedback that solo study can't replicate.

The hard part isn't the math — it's staying structured under pressure while narrating your thinking to an interviewer. That's a performance skill, not a knowledge skill. It develops through repetition in realistic conditions. Kasie is an AI case interview practice platform built by ex-MBB consultants that simulates realistic interviewer-led and candidate-led cases with structured feedback across six performance dimensions. Unlike practicing alone with a timer, you get real-time feedback on whether your structure, assumptions, and communication land the way you intend.

[INTERNAL LINK: free case interview practice]

Frequently Asked Questions

How long should a market sizing answer take in a case interview?

Aim for 3-5 minutes. Spend 15-20 seconds clarifying the question, about 30 seconds laying out your approach, 2-3 minutes on the estimation math (talking through each assumption), and 15-20 seconds on a sanity check. If a market sizing question is embedded within a larger case, keep it tighter — 2-3 minutes. An answer under 2 minutes usually means you're not showing enough structure, and over 6 minutes means you're losing the interviewer's attention.

What's the difference between top-down and bottom-up market sizing?

Top-down starts with a large known number (like total population or GDP) and applies successive filters to narrow it down to your target market. Bottom-up starts with a single unit (one store, one customer, one transaction) and multiplies up to the total market. Top-down tends to work better for consumer markets, while bottom-up works better for B2B markets and operational questions. The strongest candidates use both as a cross-check — research suggests this approach scores approximately 25% higher on quantitative assessments.

Do interviewers care if my market sizing estimate is accurate?

Accuracy matters less than you think — but it's not irrelevant. Being within 2-3x of the actual number is generally fine. Being off by 10x suggests a structural problem in your approach or a wildly unreasonable assumption. What interviewers primarily evaluate is your logical structure (40% of the assessment), the reasonableness of your assumptions (30%), your ability to communicate your thinking clearly (20%), and the final sanity check (10%). A well-structured answer that lands at $8 billion when the real answer is $5 billion will outscore a poorly structured answer that happens to land at $5.2 billion.

What are the most common market sizing questions in consulting interviews?

The most frequently tested categories are: consumer product markets ("how many diapers are sold in the US"), service markets ("what's the size of the US fitness industry"), "how many" estimations ("how many gas stations in the US"), geographic markets ("how many Starbucks could operate in London"), and TAM/SAM/SOM questions for new products or market entries. Market sizing questions appear as standalone exercises in roughly 30% of consulting interviews and are embedded within larger cases in another 20-25%, making them one of the most frequently tested skills overall.

How do I get better at mental math for market sizing?

Mental math for market sizing is more about simplification than calculation speed. Round aggressively — use 300 million instead of 330 million if it makes multiplication easier. Break complex multiplications into parts: 2.5 million × $35,000 becomes 2.5 × 35 = 87.5, then add back the appropriate zeros ($87.5 billion). Practice estimation-style calculations (not precise arithmetic) for 10-15 minutes daily over 2-3 weeks. Research on mental math training shows that candidates who practice structured estimation improve their speed and accuracy by 40-60% within just two weeks of daily practice. [INTERNAL LINK: case interview math]

Should I memorize market size data for case interviews?

Memorize reference points, not market sizes. Knowing that the US population is ~330 million, there are ~130 million households, US GDP is ~$28 trillion, and there are ~6 million businesses with employees gives you anchors to build from. But memorizing that "the US diaper market is $5.6 billion" is pointless — the interviewer wants to see you derive it, not recall it. The cheat sheet in this guide covers the 20-25 reference numbers that are actually useful across different market sizing questions.

Market sizing is the most learnable skill in case interviews. The structure is always the same — break it down, estimate, multiply, check. Master it in 15 practice questions, then move on to the parts of consulting interviews that actually separate candidates: [INTERNAL LINK: case interview frameworks] and [INTERNAL LINK: how to structure a case interview].