derekboyer62d1648fc9-nfxaq

Learn more

Triangulate Market Sizing With U.S. Data: CBP, QCEW, and 15–25 experts

Use top-down sizing for quick scoping and bottom-up sizing for defensible estimates, then run both and reconcile them. Anchor every input to a named source like the Census Bureau or BLS, label every assumption with its unit, and treat convergence within roughly 15% to 30% between your two estimates as the signal that your number is ready to defend.

TL;DR:

  • Running both top-down and bottom-up sizing methods and achieving convergence within 15% to 30% provides a defensible market estimate.
  • Federal data sources like Census CBP and BLS QCEW supply reliable unit counts and industry averages, especially when revenue data is suppressed.
  • Accurate assumptions for penetration, pricing, and timeframe are critical, with gaps over 30% indicating a need for assumptions revalidation.
  • Market size accuracy depends more on sourcing transparency and assumption clarity than on model complexity or effort level.
  • Investors scrutinize the credibility of assumptions and sources, favoring ranges, references, and primary research over over-precise point estimates.
Trydialectic
Stress-Test Your Investor Pitch
Dialectic exposes fragile assumptions and tough investor questions, helping founders strengthen their pitch before crucial VC meetings.
Explore Dialectic

Table of Contents

Market Sizing Methods and the TAM, SAM, SOM Framework

Market sizing is the process of estimating how much revenue or how many units a defined market represents, using named data sources and explicit math rather than a guess. Every credible sizing exercise breaks the number into three tiers, and mixing them up is the single fastest way to get laughed out of a room by an investor.

TAM (Total Addressable Market) is the theoretical ceiling: every dollar or unit that would exist if you captured 100% of demand for your category, everywhere it exists. SAM (Serviceable Addressable Market) narrows that to the slice you could actually reach given your product, geography, and channel. SOM (Serviceable Obtainable Market) is the portion you can realistically capture in a given timeframe given your team, budget, and competitive position. Each tier needs its own independent calculation, not a percentage haircut applied to the tier above it.

The core formula behind all three is Price × Quantity, usually written as P × Q. Decide upfront whether Q means annual revenue, unit sales, or active users, then hold that unit constant through every calculation. Mislabeling creates predictable damage:

  • An inflated TAM with no reachable SAM signals to sophisticated investors that you haven’t thought about distribution.
  • A SOM presented without a timeframe reads as a guess, not a forecast.
  • Switching units mid-deck (dollars in one slide, users in the next) destroys the math’s credibility instantly.

Top-Down Market Sizing: The Chain-Ratio Method

Top-down sizing starts from a macro number and narrows it with a series of filters until you arrive at your specific market. It’s the fastest way to scope an opportunity, and it’s the method most pitch decks lean on first, because the inputs are often just a Google search away.

The procedure runs in three moves:

  1. Pick a macro anchor. This is usually total industry revenue, national population, or total household spending in a category, pulled from a source like Census Bureau data or a government statistical release.
  2. Apply MECE filters. Mutually exclusive, collectively exhaustive cuts narrow the anchor down by geography, demographic, or use case, each filter removing a slice that doesn’t apply to you.
  3. Multiply by penetration and price. The remaining population gets multiplied by an estimated adoption rate and an average price point to produce a dollar figure.

Here’s a compressed version of the math. Say the U.S. pet grooming industry generates a multi-billion-dollar market annually. Filter again to the Northeast region, representing a significant portion of the U.S. population.

Top-down shines for large consumer markets, cross-country comparisons, and situations where you need a credible number in an afternoon rather than a month. It’s weaker when your filters rely on soft percentage assumptions stacked on top of each other, since each unverified filter compounds the error in the final figure.

Bottom-Up Market Sizing: Unit Economics in Practice

Bottom-up sizing starts from a single countable unit and builds upward, which is why it tends to produce numbers investors trust more than a top-down guess. You’re not slicing a big number down; you’re adding real units up.

The procedure has four steps:

  1. Define your fundamental unit. This might be a customer, a store location, a transaction, or a seat license, whatever your revenue model actually charges for.
  2. Estimate per-unit price and frequency. How much does one unit spend, and how often? A SaaS seat billed monthly behaves differently than a one-time equipment sale.
  3. Count the units. Pull real counts from Census County Business Patterns, trade directories, or your own CRM pipeline if you’re already selling.
  4. Multiply price by quantity. P × Q gives you the market size in the unit you defined.

A worked example: suppose you sell compliance software to mid-size accounting firms. CBP data shows a substantial number of U.S. establishments in the accounting services NAICS code with 20 to 99 employees, your target segment. If your annual contract value averages several thousand dollars per firm and you assume a realistic obtainable penetration over five years, you can estimate the obtainable market accordingly.

Bottom-up works best for B2B and niche markets where the buyer count is small enough to observe directly, and it’s the recommended route when your product targets a limited number of identifiable buyers you can actually list.

Choosing Top-Down, Bottom-Up, or Both

The choice usually comes down to three factors: how much reliable data exists, how complex the market structure is, and how defensible the number needs to be. A quick internal planning estimate can survive on top-down alone. A number going into an investor deck cannot.

  • Data availability drives the starting point. If unit counts exist (store locations, licensed businesses, registered firms), start bottom-up. If only aggregate industry revenue exists, start top-down.
  • Complexity favors bottom-up. Markets with multiple customer segments and different price points get distorted by top-down filters that assume uniformity.
  • Defensibility requires both. Any number facing investor or board scrutiny should be triangulated, not single-sourced.

The working rule of thumb: run both whenever you have the data to do so, and treat a gap of more than 30% between the two estimates as a signal to audit your assumptions, not to average the two numbers together and move on. The biggest gaps almost always trace back to penetration rate assumptions or price point estimates, since those are the inputs least grounded in observable data.

US Federal Data Sources for Market Sizing

Federal statistics are free, methodologically transparent, and updated on a predictable schedule, which makes them more defensible in an investor conversation than a proprietary analyst report you can’t show your math on. Four sources cover most sizing needs:

The cookbook: find your NAICS code, pull establishment and employment counts from CBP, then check whether the revenue figure for that cell is suppressed. Census suppresses revenue data at the county or fine NAICS level to protect firm confidentiality in categories with few businesses.

When revenue is suppressed: pull the QCEW employment count for that NAICS code and multiply by an industry average revenue-per-employee figure from SUSB or a trade publication. If your NAICS cell shows 340 employees and comparable firms in that industry generate roughly $180,000 in revenue per employee, your proxy estimate lands near $61 million, a defensible substitute for a number the Census Bureau won’t publish directly.

Three data sources converging into an estimate

How to Triangulate and Defend Your Market Size

A single method, run once, is an estimate. A number built from three independent layers is a finding you can defend in front of a board.

The three layers are desk research (industry reports, government data), internal or client data (your own CRM, sales pipeline, or client records if you have them), and primary research, meaning direct expert interviews. Primary interviews are usually the layer that moves a rough estimate into a number you’d actually bet money on.

Run the five-step routine in order:

  1. Build your top-down estimate from a macro anchor and filters.
  2. Build your bottom-up estimate from unit counts and price.
  3. Compare the two and calculate the percentage gap.
  4. If the gap exceeds roughly 30%, interview 15 to 25 people across customer, channel, ex-vendor, and analyst categories to pressure-test the assumption driving the largest divergence.
  5. Reconcile into a final range, not a single point estimate, and document which assumptions moved the number most.

Pro Tip: Keep a running log of every assumption’s source and date as you build, not after. Reconstructing where a penetration rate came from six weeks later, right before an investor meeting, wastes an afternoon you don’t have.

Any appendix or slide backing this number should include the date of every input, the named source for each figure, the assumptions behind any percentage you invented, and a sensitivity range showing how the total moves if a key driver shifts by 20%.

A Worked Example: Sizing a B2B Subscription Tool

Scope: a project management tool for independent architecture firms in the United States, priced at $600 per year per firm, measured as annual recurring revenue for 2026.

Bottom-up build:

  1. Pull the NAICS code for architectural services from CBP: roughly 92,000 establishments with fewer than 20 employees.
  2. Assume 35% of these firms currently pay for any project management software, based on a rough industry adoption estimate.
  3. Multiply: 92,000 × adoption rate × $600 = a multi-million-dollar SAM.
  4. Apply a five-year obtainable penetration of 8%: SAM × obtainable penetration rate = an obtainable market in the low millions.

Top-down build:

Both methods land near $1.5 million. When estimates land this close, present the range ($1.4 to $1.6 million) rather than a false-precision single figure.

Common Market Sizing Pitfalls and a Documentation Checklist

Most sizing numbers fail for the same handful of reasons, and nearly all of them are fixable with better labeling rather than better math.

  • Category creep: quietly widening the market definition mid-analysis to make the number look bigger.
  • Missing units and dates: a figure with no stated year or currency unit is unusable six months later.
  • Unlabeled assumptions: any penetration rate, price point, or growth figure that isn’t sourced or flagged as an estimate.
  • No sensitivity check: presenting one number instead of a best case and worst case built from ±20% swings on your top two assumptions.

Pro Tip: Build a one-page appendix with four columns: assumption, source, date pulled, and sensitivity range. Investors trust decks that show their work more than decks that show a bigger number.

Why Investors Stress-Test Your Market Sizing Assumptions

Investors don’t push back on your TAM slide because they doubt the math. They push back because penetration rates, unit prices, and growth assumptions are where founders most often paper over weak thinking with a confident-sounding number.

Expect direct questions on:

  • How you validated your penetration rate assumption, and against what comparable
  • Whether your unit economics hold at the SOM you’re claiming to capture in year one
  • What happens to your number if your growth rate assumption is cut in half

Adversarial review, the kind that mirrors a partner meeting rather than a supportive editing pass, surfaces exactly these fragile spots before an investor finds them first. A red-teamed board memo shows how forensic scrutiny on financial assumptions changes the narrative long before the room does.

What Accuracy to Expect and How to Scope the Effort

A quick desk-research sizing, built in an afternoon from CBP and QCEW pulls, is fine for internal planning and early-stage scoping. Expect error bands wide enough that you’re really estimating an order of magnitude, not a precise figure.

Match your effort to the decision: don’t spend three weeks triangulating a number that only needs to survive an internal roadmap conversation, and don’t ship a back-of-napkin TAM into a Series A deck.

If your fundraising timeline allows for real rehearsal, running your sizing assumptions through a structured audit, like Trydialectic’s Tactical Audit at $25 per one-off pass, or the Founder Pro plan at $49 per month for ongoing deck stress-testing, catches the fragile assumption before a partner does. For teams preparing multiple pitches across a fundraising cycle, the Venture & Studio 5 Seats plan at $149 per month extends that same adversarial rehearsal across a full team.

Practitioner’s Take: Precision Is Not the Same as Defensibility

Founders chase precision when they should be chasing defensibility. A TAM slide with a number carried to the decimal point looks impressive and means almost nothing if the penetration assumption behind it is a guess dressed up as research. Investors have seen thousands of decks with a suspiciously round $4.7 billion TAM and a footnote citing a market research firm nobody’s verified.

What actually earns trust is the opposite instinct: show your range, show your sources, and show where you’re uncertain. A founder who says “our SOM is $1.4 to $1.6 million depending on penetration rate, and here’s the CBP data and the 18 customer interviews behind it” sounds more credible than one who says “our TAM is $40 billion” and can’t explain the filter chain that got them there.

The gap between a sizing exercise that survives a partner meeting and one that collapses under the first hard question almost never comes down to the model. It comes down to whether every number in the model traces back to something a skeptical stranger could check in five minutes. That’s the actual bar, and it’s a lower one than most founders think, if they build toward it from the start instead of trying to defend a guess after the fact.

— D

Sources

FAQ

What Are the Main Types of Market Segmentation?

Market segmentation methods commonly used in sizing work include demographic, geographic, psychographic, behavioral, firmographic (for B2B), needs-based, and usage-rate segmentation. Each cuts your TAM into smaller, more addressable slices, and picking the right one depends on whether your buyer differences are driven by who they are, where they are, or how they behave.

Does McKinsey Use Market Sizing Methods?

Yes, top-tier strategy consulting firms including McKinsey routinely use both top-down and bottom-up market sizing, often in case interviews and client engagements alike. The chain-ratio, top-down approach is a staple of consulting case interviews specifically because it tests a candidate’s ability to structure an estimate from limited data.

Can You Give an Example of a Market Sizing Exercise?

A simple example: sizing the U.S. market for a $600 per year subscription tool sold to architecture firms. Pulling roughly 92,000 firms from Census CBP data, applying a 35% software adoption filter, and multiplying by price yields a SAM of several million dollars, then applying an obtainable penetration rate over five years results in a SOM in the low millions.

What Are the Main Market Sizing Techniques?

The two core techniques are top-down (chain-ratio), which narrows a macro anchor with filters, and bottom-up (unit economics), which counts individual units and multiplies by price. Best practice pairs both and checks for convergence within roughly 15% to 30% before treating the number as final.

How Long Should a Market Sizing Study Take?

A quick desk-research estimate using federal data sources can be built in a day or two for internal planning. An investor-grade study that triangulates desk research, internal data, and primary interviews typically takes two to four weeks to reach a defensible range.

Leave a Reply

Discover more from derekboyer62d1648fc9-nfxaq

Subscribe now to keep reading and get access to the full archive.

Continue reading