Validation guides
Market Sizing for a Niche B2B Startup: Reachable Accounts
A niche B2B market is not measured in people. It is measured in accounts: companies that fit, have a budget, and can be reached. Here is how to count them, with public sources, and why the reachable number matters most.
To size a niche B2B market, count accounts rather than people: the number of companies that match your ideal customer profile, multiplied by what each would realistically spend a year, then narrowed to the accounts you can actually reach. Use government business counts for the first number and your own sales evidence for the last.
This guide builds on our TAM, SAM, SOM worked example, which sized a market by counting individual professionals. For most B2B products the buyer is a company, so the unit changes.
Key takeaways
- Count accounts, not people, when a company is the buyer.
- Filter by your ICP: industry, size and situation.
- Spend per account comes from buyer conversations and competitor pricing.
- Reachable accounts is the number that decides whether the business works.
- Cite a public count for the starting number.
Why should B2B startups size markets by accounts?
Because in B2B the company buys, even if many people use the product. Counting people inflates the market and hides how few buying decisions there really are.
A tool used by ten people at one company is one sale, one contract and one budget owner. If you size by people, you count it ten times. Sizing by accounts keeps the number tied to buying decisions, which is what your sales capacity is measured against.
How do you size a niche B2B market by accounts?
Four steps: count the companies in the industry, filter to your ICP, estimate annual spend per account, and narrow to the accounts you can reach.
| Step | What you do | Where the number comes from |
|---|---|---|
| 1. Industry count | Count establishments in your industry codes | US Census County Business Patterns, by NAICS code and employment size (Census) |
| 2. ICP filter | Keep those matching size, location and situation | CBP size classes, plus your own criteria |
| 3. Spend per account | What a fitting account would pay a year | Competitor pricing, buyer interviews, pilot prices |
| 4. Reachable accounts | The share you can actually get in front of | Your channel tests: replies, meetings booked |
County Business Patterns reports "the number of establishments, employment during the week of March 12, first quarter payroll, and annual payroll" by industry and geography, which makes it a citable starting point for the first two steps.
What does an account-based sizing look like with numbers?
Here is the arithmetic with illustrative inputs. Replace each one with your own sourced figure; the structure is what matters.
| Line | Example input | Running total |
|---|---|---|
| Establishments in your industry code | 20,000 (look up yours in CBP) | 20,000 accounts |
| Share in your size range | 30% | 6,000 accounts |
| Share meeting your situation (for example, multi-state) | 50% | 3,000 accounts |
| Annual spend per account | $4,800 | $14.4M a year addressable |
| Share reachable with your channel in two years | 10% | 300 accounts, $1.44M a year |
The last line is the one to take to investors and to your own planning. It is also the one your early outreach can test directly: if you cannot book meetings with a meaningful share of the first 50 accounts you contact, the reachable share is lower than you assumed.
How do you estimate spend per account?
Triangulate from three sources: what competitors charge accounts like these, what buyers told you they spend on the problem today, and the price accepted in your pilots.
Competitor pricing pages give a range. Interviews tell you what the problem costs today, which sets a ceiling on value. Pilot prices tell you what buyers actually paid. Where the three disagree, trust the pilot price most, and show the range. Our guide to B2B SaaS pricing covers testing willingness to pay.
Why is the reachable market the most important number?
Because it is the number your business actually runs on. A large addressable market you cannot reach produces no revenue.
Paul Graham's advice to start with "something a small number of people want a large amount" (Paul Graham) fits B2B niches well: a few hundred reachable accounts that need the product badly can support a strong early business, and are far easier to win than a vague market of millions.
What tools help with B2B market sizing?
Government business statistics for counts, competitor sites for pricing, and a structured calculator that keeps each assumption visible.
NELL's free TAM, SAM, SOM calculator walks through the same narrowing steps. DeepValidate adds sourced market sizing for a specific idea from live research, with every figure linked to where it came from.
Frequently asked questions
How do you calculate the market size for a B2B product?
Count the companies that fit your ideal customer profile, multiply by what each would realistically spend a year, then narrow to the share you can reach with your channels.
Where can I find the number of companies in an industry?
In the US, the Census Bureau's County Business Patterns reports establishment counts by industry code, geography and employment size. Other countries publish similar business statistics.
Should B2B market size count users or companies?
Companies, when the company is the buyer. Counting users inflates the market and hides how many separate buying decisions you need to win.
What is a good market size for a niche B2B startup?
One whose reachable accounts, at a realistic price, support the business you want to build. A few hundred accounts that need the product badly can be enough to start.
How accurate does B2B market sizing need to be?
Accurate enough to make a decision, with every assumption visible. The reachable share is the most uncertain input, so test it with real outreach.
Start where you are
Start from a public count. Finish with the accounts you can reach.
Sources
The example sizing uses illustrative inputs, not data. Look up the establishment counts for your own industry codes before relying on any figure.