Market Sizing Study: TAM, SAM, SOM and Growth Forecast

Questions this report answers

What is our category worth today, in euros and in units?
How much of it can we serve with our product, channels and countries?
What share is realistic within three years, and what would it take?
What growth rate is defensible for the forecast?
Why do the top-down and bottom-up estimates differ?
Which assumptions move the result most if they turn out wrong?

When you need it

Commission a sizing study when a number will be used to commit money. Typical cases: a start-up whose investors want a TAM they can check rather than a slide with a round billion on it; a manufacturer deciding whether a niche justifies a dedicated production line; a sales director testing whether a 10% country share is ambitious or trivial; and a company comparing several countries, where the same model must run on each candidate — often as one strand of a market entry study.

Sizing matters after the event, too. If revenue has plateaued, a reliable estimate of the serviceable market shows whether you have reached the ceiling of your segment or are losing share to rivals — two problems with very different cures.

A stand-alone study is the wrong choice when you also need to understand why customers buy or who the competitors are; a full market research report then includes sizing as one chapter. Nor is it worth doing for a category so new that no comparable spending exists; there, a demand test with real prospects beats any model built on proxies.

Sample table of contents

Executive summary
TAM, SAM and SOM on one page, each with a range
Market definition
product boundaries, customer types, geography, base year
Methodology
sources, both estimation routes, reconciliation
Top-down estimate
from official totals down to the target segment
Bottom-up estimate
buyers × frequency × average spend, by segment
Triangulation
comparison of routes and the chosen central estimate
Value versus volume
units, revenue and the price assumptions linking them
Historical growth
past series and CAGR
Forecast and scenarios
base, low and high cases over three to five years
Sensitivity analysis
the inputs with the largest effect on the result

Methodology and sources

Two routes, one answer

The top-down route narrows an official total step by step: population, share in the target group, share buying the category, spend per buyer. For B2B categories the totals often come from the structural business statistics used in an industry analysis. The bottom-up route counts real buyers or outlets and multiplies by observed frequency and spend. The gap between the routes is itself a finding, and we document how it was closed.

Getting the inputs right

Eurostat's online-shopping share — 73.6% across the EU in 2025 — covers individuals aged 16–74, so applying it to the World Bank's total EU population of 451.1 million inflates the base. Spending data hides a similar trap: Eurostat's 2024 real expenditure per head on food and non-alcoholic beverages is 2,200 PPS in Bulgaria and 3,100 in the Netherlands; purchasing power standards strip out price differences, so they compare volumes consumed, not euros spent.

Primary data where proxies fail

When no published series fits, a short CAWI survey measures penetration and purchase frequency, and distributor interviews establish volumes and realised prices.

Growth rates

Growth is expressed as a CAGR with its start and end years. Eurostat shows net turnover of Bulgaria's market research industry (NACE M73.2) rising from €127.4 million in 2021 to €182.8 million in 2024 — a nominal CAGR of about 12.8%. The Dutch series peaked in 2023 and then fell, a reminder that one smooth rate can hide a volatile path.

What you receive

Sizing report (PDF) with TAM, SAM, SOM, method notes and sources
Excel model with separate top-down and bottom-up tabs
Scenario sheet with base, low and high forecasts
Tornado chart showing which inputs matter most
One-page summary for a pitch deck or board paper
Review call with your finance team

How the project runs

Scoping call and market definition — 1–3 days
Source review and data collection — 1–2 weeks
Expert interviews or short survey, if needed — 1–2 weeks
Modelling and triangulation — 3–5 days
Draft and challenge session — 3–5 days
Final model and report — 1–2 days

Hypothetical example for illustration

Worked example

Example (hypothetical): a German company sells appointment-booking software to dental practices and wants to know whether the Netherlands and Bulgaria justify their own sales team. Its pitch deck multiplies the number of dentists by its list price and calls that the market.

The study rebuilds the figure. Top-down, it starts from practices listed in national business registers and removes those tied to hospital or chain systems. Bottom-up, interviews with practice managers establish how many already use booking software, what they pay and how often they switch. The Dutch SAM proves several times larger than the Bulgarian one, but most Dutch practices already have software and rarely change; in Bulgaria adoption is low, so the realistic SOM depends on first-time buyers.

The board funds one Dutch sales hire with a switching offer and tests Bulgaria through a reseller before adding headcount. The model is handed over so penetration can be updated as sales arrive.

Common mistakes and how to avoid them

Presenting TAM as the opportunity. It includes customers you cannot reach, and investors discount it at once; lead with SAM and SOM.
Multiplying units by list price. Discounts, promotions and product mix change the result; use average realised prices.
Using the wrong base population. Many shares apply to an age band or to firms above a size threshold, so match the base before multiplying.
Extrapolating one good year. A forecast built on the latest spike overshoots; use several years and state the scenario logic.
Converting at exchange rates alone. Spending power differs between countries, so check price-level indices.

Who it suits

Suitable for	Why
Start-ups raising seed or Series A funding	Investors test the logic behind each layer, not just the headline
Manufacturers assessing a niche	Shows whether volume justifies capacity
Sales and finance teams setting targets	Turns market size into realistic share goals
Private equity and corporate development	Checks a target company's market claims independently
Firms comparing EU countries	Runs one consistent model across all candidates

Frequently asked questions

What is the difference between TAM, SAM and SOM?

TAM is total category demand if every possible buyer purchased. SAM is the part your product, price position, channels and geography can serve. SOM is the share you can realistically win in a stated period. The TAM/SAM/SOM calculator and our worked sizing example show how the layers relate.

How long does a market sizing study take?

Sizing built mainly on published data takes two to three weeks. If inputs must come from expert interviews or a survey, allow four to five weeks. Multi-country work adds time, because national statistics often use different definitions and reference years that must be aligned first.

What affects the cost of a sizing project?

How well official statistics cover the category, whether primary data collection is needed, the number of countries and segments, and the depth of forecasting. A niche B2B market with no published series takes more work than a well-documented consumer category. To discuss your case, ask us for a quote.

Top-down or bottom-up: which is more accurate?

Neither on its own. Top-down estimates are quick and consistent but depend on how closely official categories match your product. Bottom-up estimates sit closer to real buying behaviour but are sensitive to sampling. Running both and explaining the difference yields a defensible range rather than one fragile number.

Is my company data kept confidential?

Yes. Sales figures and customer data shared for calibration are used only in your model, stored securely and deleted or returned when the project ends. Interview and survey respondents are anonymised in line with the GDPR, so nothing in the deliverables identifies an individual.

Can the forecast be updated later?

Yes. Each assumption sits in a single cell with its source, so when new statistics or your own sales figures arrive, the forecast recalculates. We can refresh the inputs annually, or your team can maintain the model itself after a short handover session.

Sources

Updated: 5 Nov 2026

Market sizing: TAM, SAM and SOM estimates you can defend