Calculating market size is basically answering one question with evidence: “How much could this market spend per year on what I sell?” The most defensible way to do it is to show (1) a clear market definition, (2) a transparent equation, and (3) data sources for every input.
Define the market precisely (otherwise every number is wrong)
Start by writing a one sentence definition that includes customer, geography, product scope, and time period. For example: “Commercial cleaning services purchased by businesses in Australia in 2025.” Market sizing collapses when the scope quietly expands, like mixing residential + commercial, or Australia + global.
Pick your market size unit and equation
Use annual revenue as your primary market size unless you’re selling units. The core equation for most service businesses is:
Market revenue = Number of buyers × Purchase frequency per year × Average price per purchase
If it’s contract based (common in cleaning, software, etc.), use:
Market revenue = Number of buyers × Average annual spend per buyer (ARPA)
These are the same idea, just packaged differently. Bottom up market sizing is literally “customers × revenue per customer.”
Method 1: Bottom up (most credible)
Bottom up means you build the market from real-world building blocks and show the math.
- Count the buyers (or sites)
Use official datasets where possible: business counts by industry, workplaces, facilities, households, etc. Then filter to the ones that plausibly buy your service. - Estimate annual spend per buyer
Get this from:
- Your own quotes and invoices (best)
- Competitor pricing sheets
- Tender documents and procurement portals
- Industry benchmarks or credible market reports
- Multiply and show assumptions
Example structure (you would plug in your own data):
“If there are 50,000 target sites, and the median contract is $4,000 per year, then the segment is ~$200M per year.”
This is the style of bottom up logic investors and analysts expect, because every number is challengeable and improvable.
Method 2: Top down (useful for triangulation)
Top down starts with an external “industry revenue” figure, then narrows it down to your segment. For example, a report might say the industry is $X, then you allocate by:
- Geography (Australia vs NSW)
- Segment (medical vs offices vs industrial)
- Channel (in house vs outsourced)
- Your addressable footprint
Top down is fast, but you must prove your split assumptions.
Method 3: Value theory (for new or differentiated offers)
If the market doesn’t exist cleanly yet (new product, new category), you estimate size based on the economic value you create and the share you can charge for. This method is commonly described as “value theory” in market sizing and TAM work.
Example logic: “We save a clinic 2 admin hours per week; that’s ~$X in wages; we price at 20% of savings; multiply across Y clinics.”
TAM, SAM, SOM (how to present it cleanly)
- TAM: the whole market if you captured 100% demand.
- SAM: the portion you can actually serve (your geography + capability constraints).
- SOM: the realistic share you can win in a time horizon (often 2–5 years).
What “data and in depth claims” looks like in practice (worked example using published numbers)
If we take Australia commercial cleaning as an example, IBISWorld reports industry revenue reaching about $20.1bn in 2025 and also lists a business count figure for the industry.
If you want to size your serviceable market (SAM) for, say, “medical clinics in NSW,” you’d do:
- Start with the national industry total (top down anchor) from IBISWorld.
- Apply a NSW share (using ABS state share of businesses or employment for the relevant customer industries, or your own customer distribution).
- Apply a “medical clinics” share (from procurement data, your own CRM mix, or a defensible proxy like counts of clinics × average cleaning spend).
- Cross check the outcome against at least one other reputable market estimate. For example, Expert Market Research publishes an Australia cleaning services market estimate around AUD 18.97bn in 2025 (different scope can explain differences).
The point isn’t that any single report is perfect; it’s that your final number is triangulated and your assumptions are transparent.
How to make your market size hard to argue with
Write your final result as a claim with inputs:
“Based on X buyers (source), Y% adoption (assumption), average annual spend of $Z (source), the SAM is $A per year.”
Then add a sensitivity band:
“If spend is $Z ± 20% and buyer count is X ± 10%, SAM ranges from $A_low to $A_high.”
That’s the difference between a market size that reads like marketing, and one that reads like analysis.