Bottom-up market sizing starts with real customers and scales up. You count the buyers you could serve, multiply by what each one pays per year, and the result is your market. That is why investors and operators trust it more than a slice of some giant industry report. Below you get the formula, a worked example, a validation checklist and the mistakes that quietly wreck estimates.
TL;DR
- The formula: TAM = number of potential customers × average annual revenue per customer.
- TAM, SAM, SOM: TAM is everyone who could buy. SAM is the part you can actually reach. SOM is the part you can realistically win soon.
- Validate it: test every input with customer interviews, your own pipeline and third-party data.
- Pair it with top-down: a big gap between the two numbers tells you which assumption to revisit.
- Be conservative on SOM: never assume you win 100% of your reachable market in the short term.
Table of Contents
- What bottom-up market sizing means
- Top-down vs. bottom-up at a glance
- The bottom-up formula
- How to build a bottom-up estimate, step by step
- A worked example: TAM, SAM and SOM
- How to validate your assumptions
- Common mistakes that make the number unreliable
- When to use top-down, and how to reconcile both
- Our perspective
- Put your customer data to work
- FAQ
- Sources
- Ready to Stop the Revenue Leak?
What bottom-up market sizing means
Bottom-up sizing builds the market from individual customer clusters. You identify the segments you plan to reach, then estimate how big each one is and how fast it grows. That is the description given in a World Bank guidance document, which also defines top-down sizing as the chain-ratio method: start with a whole market universe and apply filters until you reach the net market.
Karl T. Ulrich, writing about beachhead markets, calls the bottom-up approach "counting noses": you form a hypothesis about your first market and list every potential customer in it.
Andreessen Horowitz describes bottom-up sizing as one that reflects the target-customer profile, willingness to pay and how the product will be marketed and sold. Top-down TAM, by contrast, rests on total market size and an assumed share of it.
In plain terms: top-down asks "how big is the industry, and what slice can I get?" Bottom-up asks "who exactly will pay me, and how much?"
Top-down vs. bottom-up at a glance
| Bottom-up | Top-down | |
|---|---|---|
| Starting point | Individual customers or transactions | A larger market that contains yours |
| Core math | Customers × revenue per customer | Market universe × filters or share |
| Data needed | Customer counts, pricing, usage | Analyst reports, government and industry data |
| Speed | Slower, more research | Fast, rough |
| Credibility with investors | Higher, assumptions tie to your go-to-market | Lower, can look like wishful share-of-market |
| Main risk | Small assumption errors cascade | Overstating the relevant market |
Gartner describes top-down modeling as a broad first estimate built from competitor revenue, analyst forecasts, government data and industry-association data. Pear VC's view is that investors find bottom-up more convincing because its assumptions connect directly to how you define the customer, your revenue model and your go-to-market plan.
To see how top-down math works, take the World Bank's example. It multiplies 20,000,000 rural adults by 20% farmers, 70% mobile-phone access and 40% eligibility to reach 1,120,000 potential customers. Every filter is a guess about a population. Bottom-up replaces those guesses with a list of actual buyers.
The bottom-up formula
The core formula is simple. Ahrefs states it as total potential customers × average annual revenue per customer. Chase for Business gives the same formula: average revenue per customer multiplied by the total number of potential customers in the target market.
The same Chase guide has a clean small-business example: $5 average spend per visit × 3 visits a week × 52 weeks equals $780 per customer per year. With 20,000 potential customers, TAM comes to $15.6 million.
Two inputs, two questions:
- How many customers could buy? Use industry reports, census data or trusted organizations, as Ahrefs suggests.
- What does each one pay per year? Estimate from expected pricing and how long customers stay. Pear VC adds that you can build average annual revenue per customer by multiplying transaction volume by price.
If you already have customers, Semrush recommends dividing annual revenue by current customers to get your average, then multiplying by the estimated number of potential customers. Want a refresher on that input? Our guide on what revenue per customer means walks through it for business owners.
How to build a bottom-up estimate, step by step
1. Define the customer precisely
Write down who buys: industry, company size, number of users, location. Gartner advises defining target segments by industry, company size and/or number of users. "Small businesses" is not a segment. "Independent HVAC shops with 3 to 20 employees in Texas" is.
If you are still working out who that buyer is, this piece on how to identify your target audience is a good companion. Our own breakdown of customer segmentation models shows how to label and act on each segment once you have them.
2. Split segments when willingness to pay differs
Pear VC recommends segmenting customers when willingness to pay differs and keeping the core revenue model simple. A two-truck plumber and a 40-truck contractor do not pay the same. Size them separately and add the results.
3. Count the customers
Pull counts from government data, trade associations and industry reports. For physical markets, Freedonia describes applying usage assumptions to a direct indicator of consumption, such as houses, vehicles or households. That method helps most when no reliable top-line demand estimate exists.
4. Set revenue per customer
Use your own sales history, surveys or competitor pricing. Carta notes that bottom-up sizing often starts with your current or serviceable customer base, estimates revenue per customer from historical sales, surveys or competitor pricing, then scales to the wider segment. When you are comparing price points, looking at how other software categories publish theirs helps, for example dance studio software pricing, which shows how monthly fees translate into annual revenue per account.
5. Multiply, then narrow to SAM and SOM
TAM is the full opportunity. Shopify defines SAM as the portion realistically reachable and serviceable, using factors such as geography and price point. SOM is the short-term share of SAM you can capture. Next section shows the math.
6. Document every assumption
Gartner is blunt on this: document assumptions and uncertainty where reliable data is lacking. A spreadsheet with one cell per assumption and a note on where it came from is enough.
A worked example: TAM, SAM and SOM
The numbers below are a hypothetical illustration, not market data. Say you sell scheduling and follow-up software to independent HVAC shops in one region.
| Step | Assumption | Result |
|---|---|---|
| Potential customers | 4,000 shops with 3 to 20 employees | 4,000 |
| Revenue per customer | $200 per month × 12 | $2,400 per year |
| TAM | 4,000 × $2,400 | $9,600,000 |
| Reachable share | 60% can be reached through your channels | 2,400 shops |
| SAM | 2,400 × $2,400 | $5,760,000 |
| Year-one capture | 2% of SAM | 48 customers |
| SOM | 48 × $2,400 | $115,200 |
The Catalyze guide from the National Heart, Lung, and Blood Institute defines TAM as the total revenue opportunity, SAM as the reachable subsegment and SOM as the short-term portion of SAM you can realistically capture. It recommends conservative SOM assumptions that reflect business constraints rather than 100% short-term capture.
For SOM, TechCrunch's startup guide suggests forecasting through expected sales, market-share timing, competition and barriers to entry. Ask how many demos your team can run, what percent close, and how long a sales cycle takes. That gives a SOM you can defend in a meeting.
How to validate your assumptions
Bottom-up estimates are only as good as their inputs. Freedonia warns that the method can take longer than top-down and is sensitive to small assumptions that cascade through the analysis, so results should be independently validated.
A practical checklist:
- ✅ Interview buyers. Gartner recommends stakeholder interviews and sales-pipeline analysis alongside third-party secondary research.
- ✅ Check your own pipeline. What share of qualified leads actually closes, and at what price?
- ✅ Test price. Would the segment pay your assumed annual figure?
- ✅ Stress the inputs. Move customer count and price up and down 20% and watch the result.
- ✅ Write down ranges. A single number hides uncertainty. Show low, base and high cases.
Your pipeline is the best reality check you own. If you want to see which leads and customers are actually ready to buy or at risk of leaving, our customer behavior scoring guide shows how small businesses turn behavior into signals.
Common mistakes that make the number unreliable
The World Bank document says estimates get shaky when segment data is unavailable or poor, when uptake and usage are hard to forecast, when willingness to buy and competitor effects are uncertain, or when point estimates obscure uncertainty. Turn that into a checklist of things to avoid:
- A vague customer definition. If you cannot list who is in the segment, you cannot count them.
- One price for everyone. Different segments pay differently. Split them.
- Counting everyone as reachable. SAM exists for a reason. Geography, price point and channel all cut the list.
- A SOM that assumes you win it all. Be conservative.
- Ignoring churn and tenure. Revenue per customer depends on how long customers stay. Predicting and preventing churn raises that number. See our guide on how to predict customer churn for service businesses.
- Hiding the assumptions. If nobody can see the inputs, nobody can challenge them.
When to use top-down, and how to reconcile both
You do not have to pick one. Gartner suggests using top-down for a rough macro-level estimate and bottom-up to validate and challenge it with specific customer-base data.
The two numbers often disagree, and that is useful. Stanford Biodesign explains that bottom-up often reflects the current solution space, while top-down may represent the total potential opportunity for a new solution. Market Research Methods recommends presenting both because each checks the other, and treating a large gap as a signal to revisit your assumptions.
Karl Ulrich adds that top-down gives broader business context and can reveal adjacent segments, while bottom-up is usually more credible because it rests on identifiable customers.
A simple rule: start top-down to see if the market is worth chasing, build bottom-up to decide what you will actually go after, then compare. If bottom-up is a tiny fraction of top-down, ask whether your first segment is too narrow or your price too low.
Our perspective
We talk to small service businesses every day, and the pattern is the same. The market-size slide gets all the attention, while the revenue already sitting inside the customer base gets ignored.
Bottom-up sizing is the right habit because it forces you to think in customers. We would push it one step further: size the customers you already have. Which ones are ready to buy more? Which ones are quietly drifting away? Those answers move your revenue-per-customer input, and that input drives the whole formula.
That is why we built Signal Engine around customer-level signals: buy-readiness scores, early churn flags and drafted next actions that wait for your one-click approval. Better inputs make better estimates, and they make a better business.
Bernard
Put your customer data to work
A bottom-up model is only a plan until your customers prove it. Signal Engine scores every customer on buy-readiness, flags who is about to churn and drafts the next action, with everything held for one-click approval before anything reaches a real customer. All features are included in every plan, with no per-seat pricing.
See how it fits your business on our small business revenue intelligence page, compare plans on the pricing page, or run the Recurring Revenue Analyzer to see how much lifetime value you may be leaving on the table.
FAQ
How do you calculate bottom-up TAM?
Multiply the number of potential customers by the average annual revenue per customer. Estimate the customer count from industry reports, census data or trade groups, and estimate revenue from your pricing and how long customers stay. Document each input so others can challenge it.
What is the difference between top-down and bottom-up market sizing?
Top-down starts with a large market and filters it down to your slice. Bottom-up starts with individual customers and builds up. Most experts suggest running both and investigating any large gap between them.
How do TAM, SAM and SOM work with a bottom-up model?
TAM is all potential customers times annual revenue per customer. SAM narrows that to the customers you can reach and serve, based on factors like geography and price. SOM is the share of SAM you can realistically win in the near term, so keep it conservative.
How do you know if your bottom-up estimate is credible?
Validate the customer base with interviews, your own sales pipeline and third-party data. Then test the inputs with low, base and high cases. If a small change in one assumption swings the result wildly, that assumption needs more research.
Sources
- World Bank guidance document on market sizing
- Andreessen Horowitz: 16 More Startup Metrics
- Karl T. Ulrich: TAM, SAM, SOM and the Beachhead Market
- Pear VC: Market Sizing Guide
- Gartner: 4 Steps for Calculating the Potential Market Opportunity
- Ahrefs: Total Addressable Market
- Chase for Business: Determining Your TAM
- Semrush: How to Calculate TAM
- Shopify: How to Calculate Market Size
- Freedonia Group: How to Determine Market Size
- Carta: Market size, TAM, SAM and SOM
- NHLBI Catalyze: The Art of Market Sizing
- TechCrunch: How to calculate your startup's TAM, SAM and SOM
- Stanford Biodesign: Top-Down and Bottom-Up Market Sizing
- Market Research Methods: Market Sizing
Ready to Stop the Revenue Leak?
Signal Engine gives small and local businesses 31 AI-powered tools to score leads by buying intent, predict churn before it happens, auto-generate email and SMS campaigns, and recover missed calls automatically — all in one dashboard starting at $49/month.
Start your free 7-day trial — no credit card required. Setup takes 5 minutes.