A useful market estimate shows what was counted, what was excluded, where each input came from and how the result changes when assumptions change. A large number beside an industry name is insufficient. Define the buying customer, purchase unit, time period, product boundary and geography before separating theoretical opportunity from the share you assume you might reach.
This is a research and planning exercise, not investment advice or a revenue guarantee. An equation does not establish demand when its inputs are guesses. Its value is making those guesses visible so the team knows which evidence to obtain before committing more resources.
Start with a definition that prevents double-counting
Is a customer an organization, a branch or an individual user? If subscriptions are sold at company level, do not multiply the branch count by a full company subscription price. For per-user pricing, define a paying user rather than counting every employee. Specify whether spending is monthly or annual and whether it covers your offer or a much broader category.
Market & Offer Research needs those boundaries before collecting numbers. Otherwise, individually valid statistics may describe different populations: an organization count from one year, a segment share using another classification, and a price aimed at much larger buyers.
The University of Waterloo’s guide recommends presenting sources, assumptions and calculations. Market-sizing reference. In practice, that means an input register another person can inspect, not simply a bibliography at the end of a presentation.
Give every input an evidence record
Record the name, unit, value, reference year, source, relevance and whether it is an observation or an assumption. If direct evidence is unavailable and a proxy is used, explain how the proxy differs from the target market. Do not describe an adjustment as a correction without evidence for its direction and magnitude.
Three distinct uncertainties often need separate treatment: how many organizations fit, what they might spend on this offer, and whether or when they would adopt it. Hiding all three inside one “expected share” makes the model difficult to challenge and update.
A source can be credible yet unsuitable for your calculation. For example, a count of registered entities may not establish the number currently operating, while an establishment count may include multiple locations under one purchasing decision. The study must explain the match rather than relying on the reputation of the source alone.
Worked hypothetical example: an internal-request subscription tool
All numbers below are invented for instruction. They are not Saudi market facts or observed prices. Assume an initial population of 2,000 organizations and three possible combinations of eligibility and annual spending. Each organization buys one subscription; its branches are not additional subscriptions.
| Input or result | Low scenario | Middle scenario | High scenario |
|---|---|---|---|
| Assumed eligible share | 25% | 40% | 50% |
| Eligible organizations out of 2,000 | 500 | 800 | 1,000 |
| Assumed annual spend per organization | SAR 1,200 | SAR 1,800 | SAR 2,400 |
| Annual value if every eligible organization buys | SAR 600,000 | SAR 1,440,000 | SAR 2,400,000 |
The final row is conditional opportunity, not expected sales. To illustrate the distinction, add separate adoption assumptions of 5%, 10% and 15%, respectively. These produce 25, 80 and 150 subscribers, with annual subscription values of SAR 30,000, SAR 144,000 and SAR 360,000. These remain arithmetic scenarios, not forecasts or measured probabilities.
The figures assume a full subscription year for every organization and exclude joining-time ramp, cancellations and discounts. If delivery capacity is only 40 customers at the middle scenario’s price, the corresponding annualized capacity ceiling is SAR 72,000 before costs, even if the market assumption allows more customers. Show that constraint rather than silently increasing capacity to match ambition.
Examine sensitivity before polishing the chart
In the middle scenario, changing adoption alone from 10% to 5% reduces annual value from SAR 144,000 to SAR 72,000 while other inputs stay fixed. That identifies adoption evidence as consequential. Moving every assumption together produces a range but does not reveal which uncertainty deserves investigation first.
Avoid taking an arbitrary small percentage of an enormous market and calling it a realistic share. Investigate reach, sales capacity, approval cycles and delivery capability instead. Competitive Analysis can describe published alternatives and offer terms, but does not establish competitors’ actual shares or customers’ willingness to switch.
A second estimate is useful when it provides an independent way to challenge the first. If two estimates disagree, inspect definitions, dates and coverage before averaging them. Agreement is not proof either if both rely on the same unsupported input.
Specify a study you can review
Request an editable model, consistent definitions, dated original sources for real numbers, an assumption register, sensitivity checks and an explanation of what cannot be estimated confidently. Ask which missing input is most likely to change the decision, and what work would improve it.
When turning the model into a Business or Investor Presentation, preserve the scenarios and their boundaries. Compressing them into a single unqualified number changes the meaning of the research. A polished diagram should not imply that uncertain inputs are precise.
Before approval, check that:
- Customer, purchase, period and geographic units stay consistent.
- Branches, users and organizations are not counted twice.
- Observations are separate from eligibility and adoption assumptions.
- Every real market number can be traced to its source, definition and date.
- Theoretical opportunity is separate from operating capacity and financial forecasting.
- Calculations can be reproduced and the most influential assumption has a verification plan.
Frequently asked questions
Do we always need TAM, SAM and SOM?
Use them when their definitions suit the decision. Acronyms do not repair weak inputs. Sometimes a transparent estimate of one specific segment is more useful than three broad circles.
Is an existing industry report enough?
It may provide context. Check its match to your product unit, geography and year. Total industry spending is not the same as spending on your particular service.
Is the middle scenario the most likely?
Not automatically. “Middle” describes an input combination, not a statistical probability. If evidence does not support probabilities, state that clearly.
What if data is scarce?
Narrow the decision, use explicit ranges and explained proxies, and identify what requires testing. A valid conclusion may be that the estimate does not yet support a substantial commitment.
Begin with a brief that defines the decision, unit and boundaries before requesting a market number. That makes the result reviewable and easier to update.
Source
- University of Waterloo: Market Sizing + Analysis. All worked-example numbers are fictional and are not taken from this source.
