Market Research Basics
Consultants routinely size markets nobody has published data on. The skill is not finding a magic source — it is building a defensible estimate from public anchors and stating your assumptions clearly enough that a client can challenge them.
Primary versus secondary research
Secondary research uses existing sources: company filings, regulatory data, industry associations, government statistics, and paid market reports. It is fast and cheap but frequently outdated and sometimes circular, since reports often cite each other back to a single original estimate.
Primary research generates new data: expert interviews, customer surveys, site visits, and mystery shopping. It is slower and more expensive but produces the insight competitors do not have. Most engagements use secondary to build the frame and primary to test it.
Top-down and bottom-up sizing
Top-down starts from a large published figure and narrows: national market size, then relevant segment share, then addressable share. It is quick but inherits the errors of the starting number.
Bottom-up builds from units: number of potential customers × adoption rate × units each × price. It is more defensible and reveals which assumption actually drives the answer.
Do both. If the two land within roughly 20 percent, the estimate is credible. If they diverge widely, an assumption is wrong — and finding it is usually the most valuable part of the exercise.
TAM, SAM, and SOM
Total Addressable Market is everyone who could theoretically buy. Serviceable Addressable Market is the portion your product, geography, and channel can actually reach. Serviceable Obtainable Market is what you can realistically capture given competition and capacity.
Most business plans fail by presenting TAM and implying SOM. A disciplined analysis shows all three with explicit narrowing logic between them.
Interviewing experts well
Prepare a short guide, ask open questions before specific ones, and always ask the closing question: what did I not ask that I should have. Ask for ranges rather than point estimates, then ask what would move the number to each end of the range.
Triangulate: no single interview is data. Three independent sources converging on a range is a finding; one confident expert is an anecdote.
Documenting assumptions
Every estimate needs a visible assumption log: the number used, the source, the date, and the sensitivity. Clients rarely reject an estimate because it is uncertain. They reject it because they cannot see where it came from.
Example: sizing an industrial services market
Question: annual market for predictive maintenance services in a country's cement industry.
Bottom-up: 42 operating plants (regulator's registry) × average 3 production lines × typical annual maintenance spend per line of $1.8M (two supplier interviews plus one published tender) × 6–9 percent typically allocated to condition monitoring = roughly $14M to $20M per year.
Top-down: national industrial maintenance market of $2.1B (industry association) × cement share of industrial output of roughly 4 percent × condition monitoring share of 18 percent = roughly $15M.
Both methods land near $15M–$18M. The estimate is defensible, and the assumption log shows exactly which three numbers would need to change to move it.
- Secondary research builds the frame; primary research tests it.
- Always size both top-down and bottom-up, then reconcile the gap.
- TAM, SAM, and SOM are different numbers — show the narrowing logic.
- Ask experts for ranges and triangulate across at least three sources.
- Maintain a visible assumption log with source, date, and sensitivity.
Module quiz
Pass mark 70%1. Your top-down estimate is $40M and your bottom-up estimate is $12M. What should you do?
2. What does SOM represent?
3. Which practice most improves the credibility of expert interview data?