TAM SAM SOM with AI: Size Your Market Right

Most founders get market sizing wrong

Here is a pattern I see constantly: a founder puts “$50 billion TAM” on a pitch deck slide, an investor asks one follow-up question, and the whole thing collapses.

The problem is not that founders lie about their market size. The problem is they genuinely do not understand what TAM, SAM, and SOM mean. They Google a Statista report, find a big number, and paste it in.

That is not market sizing. That is decoration.

If you cannot explain exactly how you got your numbers, you do not have a market size. You have a guess. And investors can smell the difference from across the room.

What TAM, SAM, and SOM actually mean

Let me break this down with a concrete example. Say you are building a project management tool for freelance designers.

TAM: Total Addressable Market

TAM is the total revenue opportunity if you captured 100% of the market. Every possible customer, everywhere, with no constraints.

For our example: every freelance designer in the world who could potentially use a project management tool. If there are 10 million freelance designers globally and they would each pay $20/month, your TAM is $2.4 billion per year.

TAM is a theoretical ceiling. Nobody reaches their TAM. It exists to show the investor that the ocean is big enough to build a real business in.

SAM: Serviceable Addressable Market

SAM is the portion of TAM you could realistically serve with your current product and business model. This is where constraints come in: geography, language, pricing, features, distribution channels.

For our example: maybe you only support English, you are priced for mid-tier freelancers ($500-5000/month revenue), and your tool only integrates with Figma. Now you are looking at maybe 800,000 designers. At $20/month, your SAM is $192 million.

SAM shows you understand your actual playing field. It proves you know who your real customer is, not just who could theoretically use your product.

SOM: Serviceable Obtainable Market

SOM is what you can realistically capture in the next 1-3 years given your current resources, team, and go-to-market strategy.

For our example: you are a solo founder with no marketing budget. You plan to get users through Figma plugin directory, Product Hunt, and designer communities. Realistically, you might capture 2,000 paying users in year one. That is $480,000 in ARR.

SOM is your actual business plan expressed as a number. It should be ambitious but defensible. If an investor asks “how do you get to this number?” you should be able to walk through the funnel step by step.

Top-down vs bottom-up: which approach to use

There are two ways to calculate market size, and you should do both.

Top-down approach

Start with a large, known market number and narrow it down.

1. Start with an industry report (“Global project management software market: $7.5B”)
2. Filter by segment (“Freelancer segment: 15% = $1.1B”)
3. Filter by geography (“English-speaking markets: 60% = $660M”)
4. Filter by your specific niche (“Design-focused: 20% = $132M”)

The problem with top-down: every filter is an assumption. Stack five assumptions and your number could be off by 10x. It is easy to manipulate. Just change one filter percentage and you get a completely different result.

Top-down is useful for a quick sanity check. It is terrible as your only method.

Bottom-up approach

Start with your unit economics and build up.

1. How many potential customers exist? (“800,000 mid-tier freelance designers using Figma”)
2. How did you get that number? (“Figma reports 4M users, industry data says 20% are freelancers, our pricing fits the middle 40%”)
3. What would they pay? (“$20/month based on competitor pricing and 50 survey responses”)
4. What conversion rate is realistic? (“2% in year one based on comparable Figma plugins”)
5. Revenue = 800,000 x 2% x $20 x 12 = $3.84M year one

Bottom-up is harder to do but much more credible. Every number has a source. Every assumption is explicit. An investor can challenge any single number without the whole thing falling apart.

The right answer: triangulate

Do both. If your top-down and bottom-up numbers are in the same ballpark, you have something solid. If they are wildly different, one of your assumptions is wrong. That is actually valuable information.

How AI can help with TAM SAM SOM calculations

AI tools are genuinely useful for market sizing. But they have specific strengths and dangerous weaknesses.

Where AI helps

Data gathering and synthesis. AI can pull together information from multiple sources faster than you can. Ask it to find the number of freelance designers in the US, the average spend on tools, the growth rate of the freelance economy. It will synthesize multiple data points in minutes.

Scenario modeling. Give AI your assumptions and ask it to build multiple scenarios. “What if our conversion rate is 1% vs 3% vs 5%?” It can run these calculations instantly and present them clearly.

Finding comparable companies. AI is great at identifying companies in adjacent markets whose metrics you can use as benchmarks. “What SaaS tools targeting freelancers have published their user numbers or revenue?”

Challenging your assumptions. This might be the most valuable use. Tell AI your market sizing and ask it to poke holes. “What assumptions am I making that might be wrong?” A good AI will identify logical gaps you missed.

If you want to go deeper on using AI for market research, check out how AI agents can do market research for a more detailed breakdown.

Where AI is dangerous

AI hallucinates market numbers. This is the biggest risk. Ask an AI “What is the TAM for project management software?” and it will confidently give you a number. That number might be real, or it might be completely fabricated. AI models do not distinguish between facts they learned from training data and plausible-sounding numbers they generated.

Always verify AI-generated statistics with primary sources. If AI says “the global PM software market is $7.5B according to Grand View Research,” go find that actual Grand View Research report. If it does not exist, the number is fake.

AI defaults to optimistic assumptions. Ask AI to size your market and it will tend to give you the largest defensible number. It is trying to be helpful, but this creates false confidence. Always ask for conservative estimates explicitly.

AI cannot assess local market dynamics. It does not know that freelance designers in your target market have strong loyalty to existing tools, or that a regulation change is about to open up a new segment. Context matters, and AI lacks your ground-level context.

A practical workflow

Here is how I recommend using AI for market sizing:

1. Use AI for initial research to gather data points and identify sources
2. Verify every number against primary sources (industry reports, government data, company filings)
3. Build your model yourself in a spreadsheet where every assumption is explicit
4. Use AI to stress-test your assumptions and find gaps
5. Never put an AI-generated number on a pitch deck without a verified source

This fits into the broader AI-powered startup strategy approach: use AI as a research accelerator, not as a replacement for your judgment.

Good vs bad market sizing: real examples

Bad example

“The global SaaS market is $200B. We are building a SaaS product. Our TAM is $200B.”

This tells an investor nothing. It is technically true and completely useless. You might as well say “the global economy is $100T and we plan to participate in it.”

Bad example (more subtle)

“There are 50 million small businesses in the US. If 10% use our invoicing tool at $30/month, our SAM is $1.8B.”

The 10% assumption is doing all the work and it has zero justification. Why 10%? Why not 0.1%? Or 30%? This is a top-down number dressed up as analysis.

Good example

“There are 2.3 million freelance designers in the US (Bureau of Labor Statistics + Upwork Freelance Forward report). Our tool requires Figma, which has 60% penetration in this segment (Figma S-1 filing). Our pricing ($20/month) targets freelancers earning $3K-10K/month, which is roughly 40% of the market (PayPal freelancer survey). That gives us a SAM of 552,000 potential customers, or $132M annual revenue.”

Every number has a source. Every filter has a justification. An investor can disagree with any single number and the framework still holds.

Good example (bottom-up SOM)

“Comparable Figma plugins (Stark, Autoflow) acquired 5,000-15,000 paying users in their first year through the plugin directory. Our landing page converts at 4% (based on beta test). With a 30-day free trial and 25% trial-to-paid conversion, we project 2,000 paying users in year one. That is $480K ARR.”

This is defensible because it references actual comparable products and uses real conversion data.

What investors actually look for

Different stages care about different things.

Pre-seed / Seed

Investors want to see that the market is big enough to support a venture-scale outcome. They are not expecting precise numbers. They want:

• A TAM over $1B (ideally $5B+) to show the ocean is big
• A SAM that makes logical sense given your product
• A SOM that is achievable and connects to a clear go-to-market plan
• Evidence you understand your market deeply, not just the numbers

The narrative matters more than precision. Can you tell a convincing story about why this market is big and growing?

Series A

Now investors want evidence that your SOM projections were roughly right. They will compare your original market sizing to your actual traction. They want:

• Proof that customers exist and will pay (you should have this by now)
• Updated TAM/SAM based on what you have learned from real customers
• A credible path from current ARR to $10-50M over 3-5 years
• Market growth data showing tailwinds

Series B+

At this stage, your actual revenue is more important than your market size estimates. But investors still care about TAM because they need to believe you can 10x from here.

Common mistakes to avoid

Mistake 1: Confusing TAM with the market you found on Google. If you Google “project management market size,” you will find the TAM for all project management software. That is not your TAM unless you are competing with Asana, Monday, and Jira head-to-head.

Mistake 2: Using TAM when you mean SAM. When someone asks “what is your market size?” they usually want your SAM. Leading with TAM makes you look like you do not understand the question.

Mistake 3: SOM that is too conservative or too aggressive. If your SOM is $50K, investors wonder why they should bother. If it is $50M in year one with no team and no funding, you have lost credibility. Find the honest middle.

Mistake 4: Static market sizing. Markets grow. The freelance economy is growing at 15% annually. If you size your market based on today’s numbers, you are underselling the opportunity. Show the trajectory.

Mistake 5: Ignoring adjacent markets. Your initial market might be small, but adjacent markets you can expand into later matter for TAM. Slack started as a gaming company tool and expanded to all knowledge workers. That expansion potential is part of the story.

How TAM SAM SOM fits into your startup strategy

Market sizing is not a standalone exercise. It connects directly to other critical decisions.

Your lean canvas should reflect your SOM. The “Revenue Streams” and “Customer Segments” boxes should be consistent with your market sizing.

Your business model determines your SAM. Pricing, distribution channels, and geographic focus all filter your total market down to your serviceable market.

Your competitive analysis shapes how you define your market boundaries. If three well-funded competitors own a segment, that segment might not belong in your SOM even if it is in your SAM.

Market sizing done right is not a pitch deck exercise. It is a strategic tool that helps you focus.

A simple template to get started

Here is a framework you can use today:

Step 1: Define your ideal customer precisely. Not “small businesses” but “B2B SaaS companies with 10-50 employees that use Stripe for payments.”

Step 2: Count them. Use LinkedIn Sales Navigator, industry reports, government databases, or competitor user numbers.

Step 3: Calculate what they would pay. Use competitor pricing, survey data, or willingness-to-pay interviews.

Step 4: Multiply for TAM (all of them), SAM (the ones you can realistically reach), and SOM (the ones you will actually get in year one).

Step 5: Stress test. What if your customer count is half? What if they pay 30% less? Does the business still work?

Step 6: Get feedback. Show your market sizing to potential investors, advisors, or other founders. They will find the holes you missed.

Do the work

Market sizing feels tedious. It is not as fun as building the product or designing the landing page. But founders who skip this step build products for markets that do not exist, or markets too small to sustain a real business.

AI can make the research faster. Tools like Startup Skill can help you structure the analysis and catch gaps. But no tool replaces the discipline of sitting down, finding real data, and building a model you can defend.

The best founders I know can explain their market sizing from memory. Not because they memorized it, but because they did the work to truly understand their market.

Do the work. Your future fundraise depends on it.