Market Sizing: Top-Down vs. Bottom-Up Approaches

Market sizing is one of the most frequently tested skills in consulting interviews, appearing in roughly 30% of first-round cases based on our analysis of 800+ case prompts. Market sizing refers to the process of estimating the total revenue, volume, or number of users for a given product or market — using logic, not research. Getting comfortable with these questions gives you an edge because the same estimation muscles power every quantitative case you will face.

Why Market Sizing Matters Beyond the Interview

Market sizing questions serve as a proxy for structured thinking under pressure. Partners at McKinsey, BCG, and Bain have noted that the ability to size a market quickly is one of the top three skills they screen for during first-round interviews. In real engagements, consultants regularly need “back-of-the-envelope” estimates to evaluate new market entries, validate client claims, or scope out investment opportunities — often within a single meeting. Practicing market sizing cases builds this instinct.

Two Core Approaches

There are two fundamental methods for market sizing: top-down and bottom-up. Choosing the right approach depends on the data you have available and the precision required.

flowchart LR
    subgraph TD["Top-Down"]
        direction TB
        A1[Total Population/Macro Data] --> A2[Geographic Filter]
        A2 --> A3[Demographic Filter]
        A3 --> A4[Usage Pattern Filter]
        A4 --> A5[Target Market Size]
    end

    subgraph BU["Bottom-Up"]
        direction TB
        B1[Single Unit] --> B2[Unit Output]
        B2 --> B3[Unit Count]
        B3 --> B4[Channel Aggregation]
        B4 --> B5[Total Market Size]
    end

    TD -.->|Cross-Validate| BU

Top-Down Approach

The top-down method starts with a large, known number and narrows it through a series of logical filters. It works best when you can anchor on a reliable macro statistic — such as total population, GDP, or industry revenue.

Step-by-step process:

1. Identify the broadest relevant population or market figure
2. Apply segmentation filters (geography, demographics, usage patterns)
3. Estimate the adoption or penetration rate for each segment
4. Multiply by price per unit or frequency of purchase

Worked example — How many cups of coffee are sold daily in New York City?

mindmap
  root((NYC Coffee Market<br>Top-Down))
    Residents
      Population 8M
        Adults 80%<br>6.4M
          Coffee Drinkers 65%<br>4.2M
            1.5 cups/day<br>**6.3M cups**
    Visitors
      Commuters+Tourists 1.5M/day
        Coffee Drinkers 40%<br>0.6M
          1.2 cups/day<br>**0.7M cups**
    **Total ≈ 7M cups/day**

• NYC population: ~8.3 million residents (round to 8M for speed)
• Adults (18+): ~80% = 6.4M
• Coffee drinkers: ~65% of adults = 4.2M
• Average consumption: ~1.5 cups per day per drinker
• Resident demand: 4.2M x 1.5 = 6.3M cups
• Add commuters and tourists (~1.5M daily): ~40% drink coffee = 0.6M x 1.2 cups = 0.7M
• Total estimate: ~7M cups per day

Notice how each step explicitly states an assumption. In our experience coaching candidates, interviewers care more about the clarity of your logic than the precision of your final number.

Bottom-Up Approach

The bottom-up method builds from a single unit — one store, one customer, one transaction — and scales up. It works especially well for B2B markets, niche segments, or situations where you have granular unit-level intuition.

Step-by-step process:

1. Define the fundamental unit (one store, one factory, one sales rep)
2. Estimate the relevant metric for that unit (revenue, output, transactions)
3. Count the number of units in the total market
4. Multiply to get the aggregate figure

Worked example — Same coffee question, bottom-up:

mindmap
  root((NYC Coffee Market<br>Bottom-Up))
    Coffee Shops
      3,500 shops × 400 cups/day
        **1.4M cups**
    Convenience Stores
      8,000 locations × 80 cups/day
        **0.64M cups**
    Office Channel
      250K offices × 15 cups/day
        **3.75M cups**
    Home Brewing
      3M households × 0.5 cups/day
        **1.5M cups**
    **Total ≈ 7.3M cups/day**

• Estimated coffee shops in NYC: ~3,500
• Average cups sold per shop per day: ~400
• Coffee shop channel: 3,500 x 400 = 1.4M cups
• Convenience stores and delis (~8,000 locations): ~80 cups/day = 0.64M
• Office and workplace brewing (~250,000 offices): ~15 cups/day = 3.75M
• Home brewing (~3M households): ~0.5 cups/day on average = 1.5M
• Total estimate: ~7.3M cups per day

Both approaches converge around 7 million, which is a strong sanity check. When your top-down and bottom-up estimates align within 15-20%, you can present your answer with confidence.

When to Use Which Approach

Based on our analysis of market sizing cases in the ProHub library, here is a practical decision guide:

flowchart TD
    A[Market Sizing Question] --> B{Reliable macro<br>data available?}
    B -->|Yes| C{Need channel-level<br>granularity?}
    B -->|No| D{Know unit<br>economics?}

    C -->|No| E[Top-Down]
    C -->|Yes| F[Both Methods<br>Cross-Validate]

    D -->|Yes| G[Bottom-Up]
    D -->|No| F

    E --> H[Start from population<br>Apply filters]
    G --> I[Start from single unit<br>Scale up]
    F --> J[Both estimates<br>within 20% = good]

    style E fill:#e8f5e9
    style G fill:#e3f2fd
    style F fill:#fff3e0

In about 40% of cases we have reviewed, candidates benefit from presenting both approaches and triangulating the answer. This demonstrates thoroughness and builds interviewer confidence in your estimate.

Seven Best Practices for Interview Performance

1. State your approach before calculating. Announce “I’ll use a top-down approach starting with the US population” so the interviewer can follow your logic and redirect you early if needed.

2. Round aggressively. Use 330M for US population, not 331.9M. Clean numbers reduce arithmetic errors and speed up calculations. For mental math tips, see our dedicated guide.

3. Segment meaningfully. Split by geography (urban vs. rural), customer type (B2B vs. B2C), or usage intensity (heavy vs. light users). Each segment should behave differently enough to justify the split.

4. Sanity-check your answer. Compare against known benchmarks. If you estimate the US coffee market at $500B, that would be ~$1,500 per person — far too high. Catching errors like this shows maturity.

5. Label every assumption. Write “Assume 65% of adults drink coffee” rather than just jumping to 65%. This makes it easy for the interviewer to challenge specific assumptions rather than your entire framework.

6. Use ranges when uncertain. Saying “between 5M and 8M, with 6.5M as my point estimate” is more sophisticated than a single number. It signals awareness of uncertainty.

7. Know your anchor numbers. Memorize roughly 15-20 key statistics — US population (330M), global population (8B), US GDP (~$28T), number of US households (~130M), US smartphone penetration (~85%). These serve as starting points for a wide range of questions.

Common Pitfalls and How to Avoid Them

Key Takeaways

• Market sizing tests structured thinking, not trivia knowledge — clarity of approach matters more than a precise final number
• Top-down works best for large consumer markets; bottom-up excels in B2B and niche segments
• Using both approaches as a cross-check is the gold standard, especially when estimates converge within 15-20%
• Round aggressively, label every assumption, and always sanity-check your final answer against known benchmarks
• Memorize 15-20 anchor statistics (population, GDP, households) to speed up any estimation question
• Based on our case library analysis, roughly 30% of first-round interviews include a market sizing component, making this a high-ROI skill to practice

Ready to put these techniques into practice? Explore market sizing cases in our case library, or test yourself under time pressure with an AI Mock Interview that provides real-time feedback on your estimation structure and math.