Market Sizing for PMs: TAM/SAM/SOM Without the Hand-Waving

A credible TAM/SAM/SOM estimate is built bottom-up, not top-down: start from a specific customer segment you can actually count, multiply by a realistic price and purchase frequency, then apply a penetration rate benchmarked against comparable companies. Skip the "$50B market" headline slide — anchor every number to a source a skeptic could go check themselves.

Quick answer: Bottom-up market sizing means counting real customers in a defined segment, multiplying by what they'd realistically pay, and applying a defensible penetration rate — not slicing a giant industry report by a guessed percentage.

Why Top-Down Market Sizing Doesn't Survive Scrutiny

Top-down sizing starts with a huge, unverifiable number — "the global AI market is $500B" — then multiplies it by an arbitrary capture rate like 2%. It fails because neither figure connects to your actual customer, your price, or your ability to sell, so any experienced reviewer can dismantle it with one question.

The tell is always the same: a massive headline figure sourced from a research firm, followed by a suspiciously round capture rate — 1%, 2%, 5% — with no explanation of where that share comes from. Investors and executives have seen this slide hundreds of times.

The moment someone asks "how did you get to 2%?" and there's no answer, the model collapses. CB Insights, which has run a recurring post-mortem analysis of failed startups for over a decade, consistently finds "no market need" among the top handful of reasons companies shut down — roughly a third of the failures it studies cite some version of it.

A hand-wavy TAM doesn't directly cause that outcome, but it's often the first symptom. A team that never counted its actual buyers rarely built the right thing for them. Before you build a slide, ask three questions that pop the top-down bubble:

• Where did the capture percentage come from — a segment you've already validated, or a wish?
• Who exactly writes the check, and how did you find them?
• What does your own pipeline data say about realistic conversion, instead of the industry's average?

A number nobody can trace back to a customer isn't a market size — it's a hope wearing a spreadsheet.

TAM, SAM, and SOM, Defined Without the Jargon

TAM (Total Addressable Market) is total demand for your category if you owned 100% share. SAM (Serviceable Addressable Market) narrows that to the segment your business model, geography, and product can actually reach. SOM (Serviceable Obtainable Market) is the realistic revenue you can capture in a defined window, usually one to three years.

Treat the three as a funnel, not synonyms for "big number." Each layer answers a different question and gets narrowed by different constraints:

For illustration, say you're selling scheduling software to independent dental practices in North America. TAM might be "every dental practice worldwide." SAM narrows to English-speaking, cloud-ready practices you can bill and support in the US and Canada — call it 200,000 practices. SOM is the slice you could plausibly sign in three years given your current sales team and channel partners, maybe 2,000–4,000 practices.

Notice that the narrowing from SAM to SOM had nothing to do with the size of the category and everything to do with your own capacity. That's the part a top-down slide always skips.

Segmenting by job, not just firmographics, usually sharpens SAM further. Two dental practices the same size can have very different underlying jobs — one is hiring software to reduce no-shows, another to simplify insurance billing. Our complete guide to Jobs-to-Be-Done walks through why job-based segments often predict buying behavior better than industry or headcount alone.

The Bottom-Up Method: Five Steps to a Defensible SOM

Building a defensible SOM means counting a specific segment with real data, pricing it against what buyers already pay, and applying a penetration rate borrowed from comparable companies. Five sequential steps replace guesswork with numbers a skeptic can trace back to a named source.

1. Define one specific segment. Not "small businesses," but something precise enough to count: "US-based Shopify merchants doing $1M–$10M in annual GMV." MIT's Bill Aulet, in his bottom-up methodology Disciplined Entrepreneurship, calls this picking a "beachhead market" — narrow enough that word-of-mouth and a single sales motion can dominate it before you expand.
2. Count it with a named, checkable source. Use U.S. Census Bureau County Business Patterns data (searchable by NAICS code), a platform's own published merchant counts, LinkedIn Sales Navigator filters, industry-association membership rolls, or G2/Crunchbase category listings. Write the source next to the number — if you can't name where a count came from, don't use it.
3. Anchor price to what buyers already pay, not what you wish they'd pay. Pull from your own closed-won deals or published competitor pricing pages rather than an aspirational future tier. Interview data is useful here too, but stated willingness-to-pay tends to run high; see why contradictory data shows up as a say/do gap before you take a survey number at face value.
4. Multiply count × price × frequency to get your SAM revenue ceiling. This is where research synthesis earns its keep — a properly synthesized set of user interviews should confirm the segment actually buys at roughly this price and cadence, not just that they said they would in one conversation.
5. Apply a penetration curve, not a guess. Look at what comparable B2B SaaS companies actually captured in years one through three of entering a new segment — OpenView Partners' annual SaaS Benchmarks research is a reasonable public proxy — typically low single digits of SAM in year one, climbing as the category and your distribution mature.

Run both in parallel and use the gap as a sanity check. If your bottom-up SOM is 40% of a top-down TAM, something in your counting or pricing is almost certainly wrong.

From Market Size to Prioritized Opportunity

A market-sizing number alone tells you how big a segment is, not whether it's worth pursuing over another segment — that takes layering in opportunity scoring: how important the underlying job is to that segment, and how well existing options already satisfy it. Size tells you the ceiling; opportunity scoring tells you which ceiling to chase first.

Compare two segments with similar SOM dollar figures. Segment A is crowded with three well-funded incumbents whose customers rate themselves highly satisfied. Segment B is smaller on paper but nobody has solved the underlying job well. A blended, company-wide TAM slide erases that difference entirely — a bottom-up SOM per segment starts to reveal it, but only opportunity scoring makes it decision-ready.

Tony Ulwick's Outcome-Driven Innovation (ODI) methodology, developed at Strategyn, gives this a formula: opportunity score = importance + max(importance − satisfaction, 0). A job rated highly important but poorly satisfied scores high; a job that's important but already well-served scores low, no matter how large the segment behind it.

Read purely by revenue ceiling, Segment A looks like the obvious first move. Read by opportunity score, Segment C — smaller in dollars — is the one where a new entrant has the most room to win, because almost nothing on the market satisfies the job yet.

Getting an opportunity score right requires the same research discipline as the sizing itself. Running consistent interviews — the kind built from a repeatable user interview question bank that asks every respondent to rate the same outcomes on importance and satisfaction — is what makes the scores comparable across segments instead of anecdotal within one.

It also helps to know where in the customer's experience the unmet need concentrates. Mapping the customer journey alongside the opportunity score often shows that dissatisfaction clusters at one specific moment — onboarding, renewal, a handoff between tools — rather than being diffuse, which changes what you'd actually build first.

Where this gets slower than it should

Most teams never connect these two exercises. Market sizing lives in a spreadsheet built once for a board deck; opportunity scoring, if it happens at all, lives in a separate synthesis document built from a different round of interviews with a different tagging scheme. Nobody multiplies the two together, so the loudest segment in the room wins by default rather than the most underserved one.

Scaling that connection across more than a couple of segments also means synthesizing a lot more interview data consistently, which is exactly the kind of repetitive tagging work that benefits from automating parts of user research synthesis rather than doing it by hand for every new segment you test.

Prodinja's Customer Jobs workspace is built to close that specific gap. It turns raw interview notes into structured JTBD statements, scores each one with Ulwick-style importance-versus-satisfaction opportunity scoring, and maps the Forces of Progress — push, pull, anxiety, and habit — that explain why a segment hasn't already switched to a better option. Pair that opportunity score with your bottom-up SOM per segment, and you get a prioritized shortlist instead of one blended slide.

Common Mistakes That Undermine a Market-Sizing Model

The most common mistakes are counting the wrong denominator, pricing at an aspirational tier instead of a validated one, and skipping a sanity check against your own sales capacity. Each is fixable by tracing every number back to a named, checkable source before it goes on a slide.

• Conflating TAM with SAM. "Everyone who could theoretically buy" is not the same question as "who my current business model, pricing, and channel can actually reach this year."
• Pricing at the tier you wish you could charge. Anchor to closed-won deals or public competitor pricing, not the premium tier you plan to launch someday.
• One blended number across very different segments. A single company-wide TAM hides that one segment's job is urgent and underserved while another's is already well-satisfied.
• No sales-capacity check. If your model implies signing 500 net-new logos next year but your team can onboard 40, the SOM is fiction regardless of how the math looks on paper.
• Treating the model as done once. Penetration rates, competitor satisfaction, and even the segment definition itself shift; a SOM calculated once at kickoff goes stale within a couple of quarters.
• Ignoring the say/do gap in pricing research. What someone tells an interviewer they'd pay and what they actually authorize on a purchase order are frequently two different numbers.

Where to Find Real Numbers (and Which Sources to Trust)

Credible market-sizing inputs come from named, checkable sources: government data for firmographic counts, research firms for top-down category ceilings, comparable companies' benchmark reports for penetration rates, and your own interview and sales data for pricing. Blend top-down and bottom-up rather than picking a side.

None of these sources is sufficient alone. A Gartner number tells you the ceiling exists; your own CRM tells you what a real buyer in your segment actually pays. Cross-checking the two is what turns a number into a model someone can defend in a room full of skeptics.

How This Connects to Structured Opportunity Scoring, Not Just Research Synthesis

Market sizing is a research technique, but its output is only useful once it's fed into a decision — which is why the strongest teams don't treat it as a standalone slide. A SOM figure paired with an ODI-style opportunity score, RICE, or Kano input becomes an actual prioritization input rather than a static fact about the world.

The discipline is the same whether you're scoring features or scoring markets: define the unit precisely, source the number, and multiply it against something that reflects real urgency — not gut feel dressed up as a spreadsheet. A market that's twice as big but half as urgent is not automatically the better bet.

Key Takeaways

• Bottom-up beats top-down for credibility. Count a segment, price it against real deals, and apply a penetration rate — a top-down "2% of $50B" invites the one question that kills the whole slide.
• TAM, SAM, and SOM are a funnel, not synonyms. TAM is the theoretical ceiling, SAM is what your model can reach, SOM is what you can realistically win in one to three years.
• Name your source for every number. Census/NAICS data, comparable companies' benchmark reports, or your own CRM — an unsourced figure isn't defensible in front of an investor or an exec.
• Segment by job, not just firmographics. A job-based segment often narrows SAM more usefully than industry or company size alone.
• Size alone doesn't prioritize. Pair each segment's SOM with an opportunity score (importance vs. satisfaction) to see which market is actually worth chasing first — the biggest number on paper isn't always the most winnable.
• Revisit the model regularly. Penetration rates and competitive satisfaction shift, so a SOM calculated once at kickoff goes stale within a few quarters.