- Market size is estimated by two independent routes, top-down from statistics and bottom-up from a survey, and confidence comes from their convergence.
- The total market, the addressable market and the realistically obtainable market are three different figures; the business plan is built on the third.
- Segmentation divides the market into groups that differ in what they buy, why and how they can be reached; a segmentation by age and income is rarely enough.
- Useful segments are measurable, large enough, accessible and different from each other in behaviour, not only in description.
- A launch market study combines desk research, a survey of the category's consumers and, usually, a qualitative phase to understand the segments.
1. The problem: a business plan built on an assumption
Most launch plans contain, somewhere near the beginning, a market figure taken from a general report, a supplier's presentation or a back-of-the-envelope estimate. Everything that follows, sales, production, marketing budget, is built on it. If the figure is wrong by a factor of two, which happens often, the plan is wrong by the same factor. Market sizing and segmentation are the two analyses that replace the assumption with a documented estimate, and they form the core of any market study for a launch or for entry into a new market.
The two answer different questions. Market size says how much: how many units, how many buyers, what annual value, at what growth rate. Segmentation says who: which groups of buyers exist, how they differ, which are accessible and profitable for your offer. Without the first, you do not know whether it is worth it; without the second, you do not know how.
2. Three figures, not one: total, addressable and realistic market
"Market size" usually hides three distinct quantities, and confusing them is the source of the biggest planning errors.
- The total market: everyone who buys or could buy the product category, nationally. It is the figure in sector reports and the least useful for a specific company.
- The addressable market: the part of the total market your product can reach with the distribution, price and positioning you have. A premium product sold in modern urban chains has no access to shoppers at rural farmers' markets, however many there are.
- The realistically obtainable market: the share of the addressable market you can win within a given horizon, taking into account competition, starting awareness and budget. It is the only figure on which a sales plan is built.
A serious market study delivers all three, with the method and assumptions of each, and shows how one leads to the next. A study that delivers only the total market has delivered a press figure, not a basis for decision.
3. Estimating market size: top-down and bottom-up
A credible estimate uses two independent methods and compares the results.
Top-down: from existing data
Start from statistics and work down. Domestic production plus imports minus exports gives the apparent consumption of the category. Foreign trade data, by customs code, give the volume and value of imports by country of origin and their trend. Population and household income statistics allow per capita consumption and the number of households in the relevant income bracket to be calculated. Sector reports, retailer data and shelf prices complete the picture. This phase is desk research and is done before any interview.
Bottom-up: from a survey
Start from buyers and work up. A survey of the relevant population measures category penetration (what percentage buys), purchase frequency, quantity per occasion and price paid. Multiplied by the number of people or households in the population, these quantities give an estimate of the market's volume and value. The survey also brings what statistics lack: the structure by brand, the purchase channels, the reasons, and all the variables needed for segmentation.
| Method | Sources | Strengths | Limits |
|---|---|---|---|
| Top-down | Official statistics, foreign trade, sector reports, retail data | Fast, inexpensive, covers the whole market, historical series | Categories defined differently from your need, data with a lag, nothing about buyers |
| Bottom-up | Survey of the category's consumers | Real behaviour, brands, segments, channels | Over-reporting of consumption, cost, a single point in time |
The two estimates never coincide exactly. If they differ by 10-20%, the estimate is solid and is reported as a range. If they differ two-fold, one of the methods has a definition or assumption error, and the reconciliation is in itself the most instructive part of the study. Respondents tend to over-report consumption of "good" categories and under-report "bad" ones, and the survey is calibrated against statistics where they exist.
4. What segmentation is and why demographics are not enough
Segmentation divides the market into groups of buyers who differ in what they buy, why they buy and how they can be reached, so that each group can receive a suitable offer and communication. The most used segmentation, by gender, age, income and region, is also the least useful: two 35-year-old women with the same income may buy for opposite reasons, and "urban women 25-45" describes a quarter of the population, not a segment.
Useful segmentations are built on variables that predict behaviour:
- Behavioural: frequency and quantity bought, channel, brand loyalty, consumption occasions. Heavy buyers of a category, usually 20-30% of buyers, often account for 60-70% of volume.
- Needs and benefits: what the buyer looks for in the category, price, quality, convenience, status, safety. This is the segmentation that leads directly to positioning.
- Attitudinal and lifestyle: values, attitudes to novelty, to local or imported brands, to health. Useful for communication, harder to reach through distribution.
Demographics do not disappear: they describe the segments after they have been defined on behaviour and needs, so that they can be found and reached. The order matters: first what they do and want, then who they are.
5. The criteria of a segmentation that can be used
A statistical segmentation can produce any number of groups; a useful segmentation meets five conditions, and the study must check them explicitly.
- Measurable: the size and characteristics of each segment can be estimated from the survey.
- Substantial: the segment is large enough to justify a dedicated offer; a segment of 3% of the market is a footnote, not a target.
- Accessible: the segment can be reached through identifiable distribution and communication channels.
- Differentiated: segments react differently to product, price and message; if two segments respond the same way, they are one.
- Stable: the segment exists a year later too, not only in the data of one wave.
Statistically, segmentation is done through cluster analysis on the behaviour and needs variables, with solutions of 4-6 segments tested for stability. Each segment then receives a full profile: size, demographics, brands used, channels, attitudes, and, most important for the decision, an estimate of its value: how many they are, how much they buy, how much they could buy from you.
6. The design of a launch market study
A complete study has three phases, and each prepares the next.
Phase 1, desk research, two to three weeks: top-down market size, trend, imports, competitors, prices, channels. It also produces the hypotheses about segments, from reports and from the company's internal data, where they exist.
Phase 2, qualitative, optional but recommended when the category is new to the company: 4-6 focus groups or 15-20 interviews with buyers of the category, to understand needs, occasions, language and choice criteria. The result is the correct list of variables for the questionnaire.
Phase 3, the survey: among the population relevant to the category, with a sample of 800-1,500 respondents, sized so that each resulting segment has at least 100-150 people for profiling. The method depends on the population: CAPI for general consumption categories, CAWI for urban, connected categories. The questionnaire measures penetration, frequency, quantity, price, brands, channels, needs, attitudes and, if the product is defined, the reaction to the concept and purchase intention at a price.
7. What the study delivers and how it is used
The final report contains: the estimate of the total, addressable and realistic market, with the two methods and their reconciliation; the structure of the market by brand, channel and region; the segments, with profile, size and value; the recommended target segments and the reasons; and the implications for product, price, distribution and communication. The figures go directly into the business plan, the segments into the positioning strategy, and the product concept, if not yet tested, moves to the testing stage among the target segments.
The study is repeated, in reduced form, one or two years after launch: the segments are checked, the share obtained is compared with the estimate, and the differences show what changed in the market and what was wrong in the assumptions. It is the cheapest way to learn from your own plan.
8. What to ask the agency
- The two estimation methods, with the sources and assumptions of each, and their reconciliation.
- The explicit distinction between total, addressable and realistic market.
- The segmentation variables and the reason they are not only demographic.
- The sample size justified by the number of segments expected.
- Verification of segment stability and of the five usefulness criteria.
- The adjustment factors for purchase intention.
- A chapter of implications, not only findings.
Market sizing and segmentation do not guarantee the success of a launch, but they remove the most frequent cause of failure: a plan built for a market that does not exist in the assumed form. A well-made market study costs a fraction of the launch budget and is the only part of it that can be verified before the rest is spent.
See also: Market research and segmentation Data analysis and desk research Quantitative research Desk research: what you learn before fieldwork Product, concept and ad testing Strategic marketing consulting
Do you have a decision to take?
Tell us the problem, not the method. We propose the right research design and explain what it can and cannot deliver.