Exploring an idea does not mean proving it will succeed. It means organizing research that helps you decide, with less uncertainty, whether an opportunity deserves further work.

What you will get from this guide

When someone says, “I want to open a coffee shop” or “I want to build a reservation app,” they are usually describing a solution already. Before investing in premises, hiring a developer, or designing a brand, it is worth investigating the situation that the proposed solution is meant to improve. GOV.UK's discovery guidance starts with four questions: who the people are, what they are trying to do, how they handle it now, and what difficulties they encounter [R02]. The SBA adds market context, demand, saturation, location, and the prices of existing alternatives [R01].

The outcome of this guide is not a business plan or a sales promise. It is a short opportunity report containing verified facts, outstanding assumptions, and a decision about what to investigate next. This is a practical editorial sequence, not an official universal method or necessarily a linear process.

Step 1. Separate the idea from the need it is meant to address

First, write down the idea as you have it: “An app that organizes reservations for small restaurants.” Then turn it into a question that does not mention the product: “What difficulties do some restaurants experience when receiving, changing, and confirming reservations?” The distinction matters. The first statement assumes software is needed; the second leaves room to discover that some restaurants prefer a notebook, that the problem only arises at peak times, or that the biggest losses occur when customers do not show up.

A useful initial record includes the group of people, task, situation, obstacle, consequence, and what you do not yet know. Do not write “restaurants need automation” if you have not observed that need. Record instead: “We believe some establishments lose time or reservations; we do not know how often or at what cost.” The dedicated article on defining problems goes deeper into writing this record.

To make this initial record more precise, see how to identify and define a business problem.

Step 2. Turn assumptions into research questions

Create two columns. Under what we believe, write: “Establishments with more than twenty reservations a day make frequent mistakes”; “the person in charge would decide to change the current method”; “we will be able to reach and speak with them.” Under what we need to find out, write observable questions: “What happened the last time a reservation was duplicated?”, “How many changes do you handle in one evening?”, “Which tools do you use, and why?” GOV.UK recommends turning unverified opinions into questions and choosing activities that produce relevant evidence [R03].

Initially, prioritize questions whose answers could change your decision to continue. You do not yet need to set thresholds for commercial experiments: formal hypothesis and test design belongs to the experimentation and validation pillar [R10]. At this exploration stage, it is enough to identify uncertainties and a responsible way of learning about them.

When you move into validation, the guide to prioritizing your riskiest hypotheses will help you decide which uncertainties to test first.

Step 3. Choose whom to study, not merely whom you know

Identify two or three contexts in which you suspect the problem occurs. For example: a restaurant with phone reservations and few employees; an establishment that combines messaging and booking platforms; or a café that does not take reservations at all. The third group matters because it may reveal that the assumed problem is not universal.

Look for participants who have experienced the situation recently. For a B2B product, try to understand separately the person doing the task, the person approving changes, and the person paying. Recruiting only friends who want to help you will bias your understanding; choosing only dissatisfied people will not represent all users either. GOV.UK's recruitment guidance emphasizes criteria tied to actual experience and a variety of profiles [R06].

For selection criteria beyond age or location alone, explore how to identify and segment potential customers.

Step 4. Design a small combination of methods

There is no magic number of interviews or mandatory mix of techniques. Choose according to the question:

Question Useful starting method What it does not establish
What happened when a reservation was lost? Interview about a recent incident How many establishments experience it
Where do management errors occur? Authorized contextual observation That a different solution would fix the problem
How many establishments operate in the area? Industry statistics and registries How many would buy a new product
What alternatives are available? Review of competitors and substitutes That their users are dissatisfied
How does demand vary across the year? Historical series and context-specific conversations A guaranteed sales forecast

You can begin with a small round, analyze it, and adjust the next one. GOV.UK recommends research cycles in which questions become more specific as findings emerge [R03]. The methods are explained in the articles on interviews, observation, and secondary data.

When the question calls for recent experiences and decisions, use the guide to customer discovery interviews.

If you need to compare interview findings with observation or records, see research methods beyond interviews.

Step 5. Investigate how the problem is handled today

Before asking whether people would adopt your proposal, examine their current alternatives: notebooks, spreadsheets, telephones, apps, messages, hiring another person, or not taking reservations. Document the value each alternative provides, the effort it requires, when it breaks down, and why it remains acceptable. Do not dismiss a manual method merely because it is manual: at a particular workload it may be sufficient and less expensive.

Complement conversations with a small competitor matrix: alternative, who it works for, observed advantage, verified limitation, and source. The SBA includes competitive analysis in the research needed to identify opportunities and barriers to entry [R01].

For an evidence-based comparison, use the guide to analyzing competitors and alternatives.

Step 6. Estimate the scale without inventing precision

Investigate the approximate universe of potential buyers and the fraction you could reach. If you estimate there are 800 establishments in a city, do not treat them as 800 customers. Separate those that take reservations, those facing meaningful difficulties, those you can reach, and those with purchasing authority. Record where each number came from, the period it covers, and how reliable it is.

Distinguish demand related to the problem from demand for your solution. A common need might be met by free alternatives; a large market may be inaccessible to a new business. Consider location, seasonality, regulatory restrictions, and frequency of use without presenting projections as facts [R01].

The estimation methods, including assumptions and seasonality, are developed in how to estimate market size and demand.

Step 7. Look for cases that contradict your intuition

Deliberately record establishments where reservations are handled successfully, workers who do not perceive the problem, and users who prefer their existing method. Ask what switching tools would involve: training, data migration, service interruptions, trust, or costs. Do not retroactively rewrite your original research question to make every finding look supportive.

Classify each finding as direct observation, participant statement, secondary data, inference, or assumption. “Four people mentioned errors” may be an accurate record; “most restaurants need the app” is not supported by those four conversations. GOV.UK recommends analyzing observations and reviewing them collaboratively to reduce individual interpretation bias [R07].

To assess contradictions without overinterpreting small samples, see how to evaluate evidence and avoid bias.

Step 8. Synthesize your findings and make a provisional decision

Complete a one-page record:

Field What to record
Problem Specific situation, without naming the proposed product
Groups People and contexts where it was observed; groups where it did not occur
Evidence Facts, records, sources, limitations, and dates
Alternatives What people use today, and why
Market Estimated universe, reachable segments, and assumptions
Uncertainties What could make the opportunity unviable and how to learn more
Next decision Move on to design, investigate further, or reframe the idea

Do not force the decision into an artificial “pass/fail.” It is legitimate to proceed with acknowledged uncertainties, investigate one more issue, or abandon an idea with a weak foundation. The signal that you can move forward is the ability to explain who experiences a meaningful problem, under what circumstances, and with what evidence—not that you have already proved your proposed solution is right.

If you need to turn this report into a reasoned decision, go deeper with how to assess a business opportunity.

Integrated example: a lunch delivery service

You begin with “I want to sell healthy lunches by subscription.” Interviews and observations among workers in one area reveal that some buy lunch several days a week; others cook at home and are satisfied; still others reject subscriptions because their work schedules change. Businesses in the area already offer delivery menus, but some buyers report delays at peak times. The research does not automatically confirm that subscriptions are the answer. It suggests investigating a specific segment, observing meal schedules, and comparing the actual reliability of alternatives. The next sensible move may be to design several different offers, not immediately hire a kitchen.

Mistakes to avoid

Researching only people you know; asking for opinions about a solution before understanding the experience; mistaking a small survey for a market measurement; presenting numbers without sources; choosing competitors solely because they look superficially similar; and continuing to research without recognizing that more information of the same kind would no longer change your decision.

What you should have when you finish

An opportunity report covering the problem, observed segments, alternatives, evidence from different sources, estimated market size and accessibility, contradictions, uncertainties, and the next step. If the problem is still unclear, redefine it. If you can already describe it well, the following articles will help you examine customers, markets, and the quality of evidence in greater depth before moving to business model design.

For more detail, see how to build a business model canvas step by step.

Main sources: [R01] SBA; [R02] GOV.UK, discovery research; [R03] GOV.UK, planning research; [R06] GOV.UK, recruitment; [R07] GOV.UK, analysis; [R10] Strategyzer. The examples are hypothetical, and the steps are an original editorial synthesis.

Bibliography and references