Market size answers one question: if everything goes well, how big can this get? For a new product there is no historical sales data, so you estimate from what you can count. This guide shows the bottom-up method, which investors trust more than big industry figures, using free US government data and a worked example.
Top-down vs bottom-up
| Top-down | Bottom-up | |
|---|---|---|
| Starts from | A published industry total | A count of customers you can reach |
| Method | Take a percentage of the total | Customers × price per year |
| Weakness | “1% of a $50B market” proves nothing | Takes longer; needs a real customer definition |
| Use for | A sanity check | Your actual estimate |
Investors discount top-down numbers because they skip the hard part: who exactly buys, and how many of them there are.
The formula
Market size = number of potential customers × average revenue per customer per year. Most founders then narrow it in three layers:
- TAM (total addressable market): everyone who has the problem.
- SAM (serviceable available market): those you can actually serve, given your product, price and geography.
- SOM (serviceable obtainable market): the share you can realistically win in the next few years.
Free US data sources for counting customers
| Source | What it counts | Good for |
|---|---|---|
| Census Bureau: County Business Patterns | Businesses with employees, by industry (NAICS code), county and size | B2B products sold to a type of business |
| Census Bureau: Nonemployer Statistics | Businesses with no employees, by industry | Products for freelancers and solo operators |
| Census Bureau: American Community Survey | Households and people by age, income, location | Consumer products |
| Bureau of Labor Statistics | Employment by occupation and industry | Tools sold per professional (e.g. per bookkeeper) |
| Trade associations | Members, surveys, industry spend | Niche industries |
| Keyword search volume | How many people search for the problem monthly | Checking demand is real |
Census tables are organized by NAICS industry codes. Find the code for your customer's industry first (the Census NAICS search does this), then pull the count.
Worked example
Illustrative figures only. Look up the current Census numbers for your own industry; these round numbers show the method, not real counts.
Product: scheduling software for independent dental practices in the United States, priced at $150 per month.
- Count customers. County Business Patterns lists establishments under the NAICS code for offices of dentists. Suppose it shows about 120,000 establishments.
- Narrow to your customer. You serve independent practices with fewer than ten staff. If the size breakdown shows roughly two-thirds are that size, that is about 80,000 practices.
- Apply the price. $150 × 12 = $1,800 per year. SAM ≈ 80,000 × $1,800 = $144 million a year.
- Estimate what you can win. If you believe you can win 1,000 practices in three years through your channels, SOM ≈ 1,000 × $1,800 = $1.8 million a year.
The SOM is the number that matters for your plan. It should connect to a channel: how exactly will you reach those 1,000 practices?
Check the estimate against real demand
Counting potential customers does not prove they want your product. Two cross-checks:
Proven and valuable
High volume, high CPC. Real demand; slow to rank. Look for narrower terms.
Often the sweet spot
Moderate volume, high CPC, lower difficulty. Buyers exist and the results are open.
Interest, little buying
High volume, near-zero CPC. Common for curiosity and informational topics.
Too early, or too niche
Low volume, low CPC. Check the problem searches before ruling it out.
- Search demand: if thousands of practices might need this, some of them should be searching for it. See estimating demand with search volume.
- Competitor revenue: if competitors exist, their size (headcount, traffic, funding) tells you whether the market supports businesses like yours.
Common mistakes
- Using the whole industry as your market. Nobody sells to all of it.
- Using an unrealistic price. Base it on what competitors charge.
- Counting people instead of buyers. A practice buys one subscription, not one per dentist, unless you price per seat.
- A SOM with no channel behind it. “We will get 1%” needs a how.
Once the numbers hold, check the rest of the business with how to know if a business idea is viable.