# How to Size Your Market Bottom-Up Before You Build

> Size your market bottom-up before you build: count real buyers, filter to who you can reach, anchor price on today's workaround, then do the math.

---
url: "https://earlyhunt.com/studio/how-to-size-your-market-bottom-up-before-you-build"
markdown: "https://earlyhunt.com/studio/how-to-size-your-market-bottom-up-before-you-build.md"
type: studio_post
title: How to Size Your Market Bottom-Up Before You Build
slug: how-to-size-your-market-bottom-up-before-you-build
published: "2026-10-26T04:50:02.307Z"
updated: "2026-10-26T04:50:12.648Z"
---

## Summary

Size your market bottom-up before you build: count real buyers, filter to who you can reach, anchor price on today's workaround, then do the math.

## Article

Bottom-up market sizing means counting the buyers you could actually name and reach, then multiplying by what each one would realistically pay you per year. Skip the "1% of a $40 billion market" slide. Pick one buyer segment, count it from public sources like government business statistics, cut that count down to the people who fit and who you can reach, anchor the price on what they spend on their current workaround, and work it out as a low, base and high range. Then do one more sum: how many customers do you need for the income you want, and is that a sane share of the reachable number? For most early founders this takes an afternoon or two, and it tells you whether the idea can pay you before you build it.

This guide is for founders who are pre-launch or in their first few months, deciding what to build and for whom. The startup and every number in the worked example are made up. Plug in your own.

## What is bottom-up market sizing, and why not just take 1% of a big number?

Top-down sizing takes a huge figure from a report and claims a small slice of it. It feels safe because 1% sounds modest, but 1% of a number you did not build tells you nothing about who will pay you. You cannot email "the market". You can email a bakery owner in Leeds.

Bottom-up market sizing flips it. You start from a single buyer and build up:

- **How many buyers like this exist?** A count you can trace to a source.
- **How many of them actually fit?** Right size, right workflow, right problem.
- **How many can you reach in the next year or two?** With your channels, your language, your country.
- **What would one of them pay per year?** Based on what they spend today, not what you hope.

Multiply the buyers you can reach by the yearly price and you get a reachable revenue number. It is smaller than the top-down figure, and it is far more useful, because every part of it is something you can check, argue with and improve.

## What do TAM, SAM and SOM mean in plain words?

You will see these three letters in every guide on how to estimate market size for a startup, so here they are in one line each:

- **TAM (total addressable market):** everyone in the world who could, in theory, buy this kind of product.
- **SAM (serviceable available market):** the part of that your product could actually serve, given what it does, where it works and what language it speaks.
- **SOM (serviceable obtainable market):** the part of SAM you could realistically win in the next few years.

Early founders often burn hours polishing TAM and skip the number that matters. This guide uses a simpler idea: your **reachable market**, the buyers who fit and who you can get in front of in the next 12 to 24 months, times what they would pay. It sits between SAM and SOM, and you can build it from facts.

## Who exactly is your buyer, and which segment do you start with?

Most bad estimates go wrong here, before any counting starts. Two rules:

1. **Count who pays, not who uses.** If your tool is used by ten bakers on a shift but paid for by one owner, the unit you count is the business, not the bakers. Users matter for the product. Buyers matter for the market size.
2. **Pick one starting segment.** "Small food businesses" is not a segment. "Independent bakeries with between 2 and 49 staff that take wholesale or pre-orders, in the UK" is. You can widen later. You cannot count a cloud.

Write the segment as one sentence with four parts: **type of business or person**, **size**, **the trigger that makes them need you**, and **geography**. If you already have a positioning draft, reuse its wording. If you do not, the exercise in [writing a one-page positioning brief](https://indiehunt.io/studio/how-to-write-a-one-page-positioning-brief-before-homepage) forces the same choices.

If you have two or three possible segments, size each one. The comparison is often the most useful output.

## Where do you find a real count of buyers?

You want a number you can trace to a source, plus a note on what it misses. Roughly from strongest to weakest:

### Government business statistics

If your buyer is a business, start here. Official statistics are free, cut by industry code and size band, and clearly defined.

- **United States:** the Census Bureau's [County Business Patterns](https://www.census.gov/programs-surveys/cbp.html) gives the number of establishments with paid employees by industry and employment size, down to county and ZIP code level. Its sister series, [Statistics of U.S. Businesses](https://www.census.gov/programs-surveys/susb.html), counts firms and establishments by enterprise size. Businesses with no employees sit in a separate series, [Nonemployer Statistics](https://www.census.gov/programs-surveys/nonemployer-statistics.html). If your buyers are solo operators, that is the one you need, because the other two leave them out.
- **Industry codes:** all of the US series are cut by industry code, so look yours up in the [NAICS code search](https://www.census.gov/naics/) first. Expect your niche to be buried inside a broader code. That is normal.
- **United Kingdom:** the ONS dataset [UK business: activity, size and location](https://www.ons.gov.uk/businessindustryandtrade/business/activitysizeandlocation/datasets/ukbusinessactivitysizeandlocation) counts enterprises and local units by SIC code and employment size band, from a yearly snapshot of the business register. For a list of actual named companies, the [Companies House advanced search](https://find-and-update.company-information.service.gov.uk/advanced-search) lets you filter by SIC code and company status. It only covers registered companies, so sole traders will be missing.
- **European Union:** Eurostat's [structural business statistics](https://ec.europa.eu/eurostat/web/structural-business-statistics) break businesses down by activity and size class, country by country.

Read the definitions first. "Enterprise" or "firm" (the business) and "local unit" or "establishment" (each site) are different counts. A bakery with four shops is one buyer and four sites, so if you sell one account per business, count enterprises or firms.

### Other counts that help

- **Professional associations and licensing bodies.** Many publish member or registrant numbers, which helps for professions like physios or electricians. Membership is usually only part of the profession.
- **Marketplace and directory counts.** Listings in your category on a booking site or app store, or installs on a related plugin. These show people who already buy software for this job.
- **LinkedIn search result counts.** Useful as a rough cross-check on job titles. Treat them loosely: profiles go stale, titles vary, and one company can show up many times.
- **Job posts that mention the job.** A business hiring someone to "manage the production schedule" is telling you the work exists and costs money.
- **Subreddits, forums and groups.** Weak signals full of lurkers and people who left years ago. Use them to find buyers to talk to, not to count them.

Log every number with its source link, the date you checked it, and what it includes or leaves out, so you can answer "where is that from?" later.

![earlyhunt-2026-10-26-inline-1.jpg](https://txmhk1zrnc.ufs.sh/f/xSkWTCqmKWx97e9V5aghzrVakeRPoULtsl3A46QO2cyG5Ejw)

## What will a buyer realistically pay per year?

This is an input for the estimate, not your pricing decision, which can come much later. You only need a defensible number per buyer per year, and the best anchor is what they already spend on the problem today.

Look for the current workaround and put a cost on it:

- **A tool they already pay for.** If they use a general tool for this job, its price is a ceiling you can see.
- **Staff or owner time.** Hours per week spent on the spreadsheet, the whiteboard or the phone, times a rough hourly cost. Do not pretend you will capture all of it. You might capture a slice.
- **A contractor or agency.** What they pay someone outside to handle it.
- **Waste or mistakes.** Only if they can name it. "We bin a tray of stock most Mondays" is a cost. "Better visibility" is not.

The fastest way to get these numbers is to ask. Talking to five target buyers about what they spend today, as in [validating an idea with five conversations](https://sidehunt.io/studio/how-to-validate-a-side-project-idea-with-five-conversations), gives you a better price anchor than any report. Ask about the past ("what did you spend on this last month?"), never "would you pay X?". People are polite about the future.

## What does a worked bottom-up estimate look like?

Here is a made-up example. **Ovenmark** is an imaginary tool that helps independent bakeries plan daily production from wholesale and pre-orders. The founder is in the UK and can only sell in English for now. Every number below is invented to show the method. None of it is real data.

Step

What you do

Made-up number

Where it would come from

1. Start count

UK bakery businesses with 2 to 49 employees under the relevant SIC code

2,700 enterprises

ONS business counts by SIC code and size band

2. Fit filter

Keep only those taking wholesale or pre-orders, so planning is a real daily job

50%, so 1,350

Your own check of 40 to 50 bakery websites and listings

3. Reach filter

Keep only those you can realistically get in front of in 24 months

40%, so 540

Your honest view of your channels: local groups, a trade show, a supplier partner

4. Price per year

Anchored on the hours the owner spends on the production spreadsheet each week

£480 per year

Five conversations with bakery owners

5. Reachable revenue

540 x £480

£259,200 per year

The sum above

Every step is a guess, so run the same sum as a low, base and high case:

Case

Fit share

Reach share

Reachable buyers

Price per year

Reachable revenue

Low

40%

25%

270

£360

£97,200

Base

50%

40%

540

£480

£259,200

High

60%

55%

891

£600

£534,600

### The back-calculation: how many customers do you need?

Now turn it around. Say the founder wants the product to bring in £60,000 a year before it is worth going full time. At the base price of £480 per year, that is 125 paying bakeries.

- **Base case:** 125 out of 540 reachable buyers, about 23%.
- **Low case:** at £360 that becomes 167 customers out of 270, about 62%.
- **High case:** at £600 that becomes 100 customers out of 891, about 11%.

This is the line that matters. Not "the market is £259,200" but "I need roughly one in four of the bakeries I can reach to pay me". That is a lot for a first product, and the low case looks close to impossible. In an afternoon, Ovenmark's founder has learned the idea only works with a wider segment or a much cheaper way to reach more bakeries.

## How do you sanity-check the number?

A bottom-up estimate is only as good as its weakest step. Three cheap checks catch most mistakes:

1. **Glance at the players who already sell here.** If two or three existing tools serve your segment, their public signals (how many customers they say they have, how many locations they list, how big their team looks) should be in the same rough world as your numbers. If your reachable market is 540 bakeries and a small rival claims thousands of UK bakery customers, one of you is counting something different. You do not need a full study for this. If you want one, see [how to do a competitor teardown before you build](https://earlyhunt.com/studio/how-to-do-a-competitor-teardown-before-you-build).
2. **Ask five target buyers what they spend today.** If three of five say "nothing, it is just me and a notebook, it takes ten minutes", your price anchor is wrong and so is your fit filter.
3. **Count a small sample by hand.** Check twenty businesses from one town one by one. What share actually fits, and what share could you contact? Your fit and reach percentages should come from this, not a feeling.

![earlyhunt-2026-10-26-inline-2.jpg](https://txmhk1zrnc.ufs.sh/f/xSkWTCqmKWx9TEDTdwZGjqPRb1rifnSW7TtzXYp0eG5ma38Q)

## How do you turn a market size into a decision?

The point of sizing a market before building is to decide something. Four decisions, in order:

### Go, widen, or drop

Use the back-calculation, not the headline number. A rough rule I find useful: if your target income needs more than about one in ten of your reachable buyers in the base case, treat that as a warning. It does not mean stop. It means you need either more reachable buyers (a wider segment, a second country, a better channel) or a stronger price anchor before you commit months of work. If even the high case needs a huge share, drop it or rethink the buyer.

### Bootstrapped or funded

For a bootstrapped product, a small reachable market can be completely fine. Needing 125 customers out of several thousand reachable buyers is a calm, buildable business, even if nobody calls it big. A funded path needs a believable route to a much larger outcome, so you would have to show how the first segment opens up adjacent ones. If the only story is "these 540 bakeries", raise less or not at all.

### Which segment to start with

If you sized two or three segments, compare reachable buyers, not total buyers. A segment of 600 you can reach through one trade association often beats a segment of 20,000 you can only reach with paid ads. Start where count x reachability is best, then widen.

### When small is fine and when to widen

Small is fine when the price anchor is strong, buyers stay for years and you can reach them cheaply. Widen when the back-calculation needs an unrealistic share, or when your reach depends on one channel you do not control.

## What traps make a market look bigger than it is?

- **Counting users instead of buyers.** Ten staff using a tool is still one invoice.
- **Double counting.** Adding sites to businesses, or adding two overlapping segments without removing the overlap.
- **Assuming 100% reachability.** Most buyers who fit will never hear of you in year one.
- **Using global numbers for a local product.** If you can only sell in one country or one language, a worldwide count is fiction. Size the market you can actually serve.
- **Trusting paid report headlines.** A press release saying a market "will reach $X billion by 2030" usually has a vague definition and a forecast behind it. Read the definition, or leave it out.
- **Spending a week on it.** Timebox it to two days. A rough range you act on beats a perfect spreadsheet you polish instead of talking to buyers.

## A 2-day market sizing plan

When

What you do

What you end with

Day 1 morning

Write the one-sentence buyer for up to three segments. Find the industry codes.

Segment sentences and codes

Day 1 afternoon

Pull counts from official statistics. Add one or two other counts as cross-checks. Log source, date and what each count leaves out.

A start count per segment, with sources

Day 2 morning

Hand-check a sample of twenty businesses to set fit and reach shares. Glance at two or three existing players' public signals.

Fit and reach percentages you can defend

Day 2 afternoon

Set a price anchor from the current workaround (book the five buyer conversations if you have not had them). Build the low, base and high table and the back-calculation.

Reachable revenue range and "customers needed" share

End of day 2

Write three lines: go, widen or drop; which segment first; which assumption is weakest.

A decision and a list of things to check next

## Market sizing FAQ

### What is the difference between TAM, SAM and SOM?

TAM is everyone who could in theory buy this kind of product, SAM is the part your product can actually serve, and SOM is the part of SAM you can realistically win in the next few years. Early on, a bottom-up reachable market is more useful than polishing all three.

### How big does a market need to be for a bootstrapped SaaS?

There is no single number, so reason backwards from your own target. Decide the yearly revenue that would make it worth your time, divide by a realistic price per customer per year, and compare the result with your reachable buyer count. If you would need a small share of the reachable market, it is big enough. If you would need most of it, it is not, at least not yet.

### What if I can't find data on my niche?

Go one level up and filter down. Take the count for the broader industry code that contains your niche, then hand-check twenty to fifty businesses to estimate the share that fits. If the buyer is a person, start from professional bodies, licensing registers or job titles. Note that the number is a proxy and widen your range to match.

### Should I put market size on my pitch deck?

If you are raising, investors will expect it, and a sourced bottom-up number beats a top-down headline. Show the count, filters and price anchor in one line. If you are not raising, the estimate is for you.

When you are ready to put an early product in front of people who like trying new things, [EarlyHunt](https://earlyhunt.com/) is a good place to start.

## Links

- Page (HTML): https://earlyhunt.com/studio/how-to-size-your-market-bottom-up-before-you-build
- AI-friendly Markdown: https://earlyhunt.com/studio/how-to-size-your-market-bottom-up-before-you-build.md

## Explore

- [Studio](https://earlyhunt.com/studio)
- [Blog](https://earlyhunt.com/blog)
- [Launch Guide](https://earlyhunt.com/launch-guide)

_HTML version: https://earlyhunt.com/studio/how-to-size-your-market-bottom-up-before-you-build_