Market research: do it yourself, step by step

Quick answer. You can run market research yourself in seven steps: frame the question, size the market in TAM SAM SOM, map the competition, build personas, run a field survey, set a price, then synthesize. The key rule: cross-check a top-down estimate (INSEE sources) with a bottom-up estimate (volume times price), using free public data.

Many founders confuse market research with a gut feeling validated by three friends. The real question is not whether people like the idea, but whether it addresses a large enough market, at a sustainable price, against identified competition. Serious market research answers these questions with figures, not impressions.

The good news: you can run it yourself, without an expensive research firm, using free and reliable sources. This article sets out the method step by step, as we see it work in the start-up files we support. The end goal is clear: produce quantified assumptions that hold up and that will feed your financial forecast.

Market research done in-house remains valid if it is structured and quantified. It helps you decide on a project, convince a funder and set realistic targets.

Here is the full procedure, in order:

1. Define the research question and target.
2. Size the market (TAM SAM SOM).
3. Map the competition.
4. Build the personas.
5. Run the field survey (questionnaire and interviews).
6. Set a price positioning.
7. Synthesize into actionable conclusions and feed the forecast.

Research starts with a precise question, not a scattergun data hunt. Ask yourself what you need to prove in order to decide whether to launch. A typical question is: in this area, are there enough customers of this profile willing to pay this price for my offer?

Then define the target with at least three criteria: customer type (consumer or business), geographic area served and precise segment. The sharper the target, the more usable the data. A fuzzy target produces fuzzy research, which is useless for the forecast.

Sizing is the quantitative core of the research. Three levels nest together: the TAM (total addressable market), the SAM (the share your offer actually serves) and the SOM (the share you can realistically capture at launch). Each level narrows the previous one using concrete criteria.

The methodological rule is to always cross two approaches. The top-down approach starts from public macro data and works down to your scope. The bottom-up approach starts from your field: number of potential customers times an estimated average basket. When both converge within the same order of magnitude, your estimate is credible.

For illustration only: if an area holds several thousand potential customers and your capacity allows you to serve a few hundred in year one, your SOM sits at that level, not at the total market level. These figures are illustrative and must be replaced with your own data.

Competitive analysis is not limited to direct rivals. Indirect competitors (substitutes, alternatives, the customer's status quo) often weigh just as much. List both families, then compare them on consistent criteria: offer, advertised price, target, channel, reputation.

Build a two-axis positioning map (for example price and service level). It visually reveals crowded zones and open spaces. Price benchmarking in particular prevents you from setting a tariff disconnected from market reality. An honest map also shows when a market is saturated, which is valuable information before investing.

A persona is a typical customer profile, described finely enough to guide your decisions. Limit yourself to two or three personas: beyond that, the analysis dilutes. Each one carries a main need, a buying context, a decision journey and a willingness to pay.

Personas connect the quantitative to the qualitative. They turn an abstract statistic into concrete customers whose motivations you understand. That understanding then shapes the offer, the sales pitch and the price. A well-built persona avoids targeting everyone, which amounts to targeting no one.

The field survey tests your assumptions against reality. It combines two parts. The quantitative part is a broadly distributed questionnaire measuring frequencies and intentions. The qualitative part relies on in-depth interviews that explain behaviour and reveal what a questionnaire cannot capture.

Mind your sample: question people representative of your target, not just your friendly circle. Beware classic biases, especially the courtesy bias where people answer what pleases. A few rigorously conducted qualitative interviews already deliver a great deal, often more than a poorly targeted mass survey.

Price is not a last-minute adjustment variable, it is a strategic choice that flows from the research. Two logics oppose and complement each other.

Arbitrage: cost-based or value-based pricing. Cost logic adds your charges and a target margin: it secures but often caps potential. Value logic starts from what the customer will pay for the perceived benefit: it frees up margin but demands a real understanding of the persona. In practice, test several price levels during the field survey, then arbitrate by crossing the cost floor with the perceived-value ceiling.

The synthesis turns the research into decisions. It fits in a few pages: market size retained, positioning, price, priority targets and identified risks. This document justifies your assumptions to a banker or a partner.

Above all, the research only has value when connected to the figures. The SOM and the chosen price become your first-year revenue assumption. That figure then feeds the financial forecast and the funding plan. Market research disconnected from the forecast remains a style exercise. Connected, it becomes a steering tool.

France offers high-quality public sources, free and usable without a subscription. The reflex to build is to always prefer a primary source over a second-hand article. Here are the main ones and their use.

For sector figures, INSEE remains the first-rank reference, and Bpifrance Création provides reliable methodological framing for founders. Crossing these two sources already covers the essentials of a credible sizing.

Some projects require adjustments. A highly specialised B2B activity will rely more on INPI open data and qualitative interviews than on mass statistics, because the target population is small and individually identifiable. Conversely, a local retail business will favour INSEE local data (catchment area, demographics, neighbourhood incomes).

An innovative project with no pre-existing market cannot rely on historical series. You then reason by analogous markets and pure bottom-up, accepting a wider margin of uncertainty. Finally, if you launch alongside a salaried job, prudent sizing matters even more.

The underestimated risk: confusing stated intention with real purchase. In questionnaires, many respondents say they are interested without ever acting. Systematically weight intentions downward and confront them with observed behaviour. Research that overstates demand produces an unworkable forecast, and that is one of the most frequent causes of a gap as early as year one.

Another point of vigilance: open data is powerful but dated. Check the reference year of each dataset and never mix different vintages in the same calculation. Finally, keep the traceability of your sources: a funder will ask where your figures come from.

Our reading. Market research is not an academic exercise separate from the figures: it is the raw material of the forecast. In the start-up files we support, the most frequent blockers come from an inflated SOM or a price set without a benchmark. Rigorous upstream research avoids months of flying blind.

We also see founders who are very strong on the product but who have never crossed top-down and bottom-up: their first-year revenue then rests on a single fragile calculation. The role of start-up support is precisely to challenge these assumptions before they become financial commitments. For more advanced structures, this logic continues into a growth and valuation strategy and monitoring through dashboards.

Recently, supporting a founder on their forecast illustrated this point: the initial study announced a very wide capturable market, but the bottom-up approach (customers actually served per week times the basket) brought the target down to a third. Choosing the prudent assumption allowed us to calibrate a realistic funding need rather than an unworkable target. This work falls within the forecast-support mission we carry out as a chartered accountant registered with the Ordre, and does not replace a personalised analysis of your situation.

Hayot Expertise tip. Build your research so that it leads to a single defensible figure: your first-year revenue. If you cannot explain it line by line to a banker, the research is not finished.

• Research question framed and target defined (customer, area, segment).
• TAM SAM SOM estimated and cross-checked top-down and bottom-up.
• Direct and indirect competitors mapped with price benchmark.
• Two or three personas with need, journey and willingness to pay.
• Field survey documented (questionnaire and interviews, sample described).
• Price positioning justified (cost and value).
• Synthesis connected to the forecast and sources traced.