Market research interviews: ask each person only what they can know

Market research interviews work when you ask each person only what they are in a position to know. They are one-on-one conversations with people who take part in a market: customers, suppliers, operators and channel partners. A customer can describe a need. A supplier can tell you why that need is hard to serve.

Those accounts answer different questions about the same market, so plan around the relationships your idea depends on. For each important assumption, name who could know, who would live with the consequences, and which other source could challenge the answer. Otherwise the study turns into a pile of loosely related opinions.

NSF's I-Corps program describes discovery as conversations with customers, partners, and other industry stakeholders. That breadth matters when a business depends on several organizations cooperating. Read NSF's description of customer discovery.

1. Draw the service as a loop of handoffs

Here is a made-up example. A founder is researching a reusable shipping-crate service for regional distributors. It works only if a customer uses the crates, someone collects them, a depot prepares them for reuse, and a supplier provides suitable stock. These are illustrative roles, not findings about a real market.

The customer may like the idea of less disposable packaging. That says nothing about whether empties can be collected at a workable point on the route, whether the depot can handle the volume, or whether the customer will accept the arrangement. (Liking reusable crates and storing a stack of empty ones by the loading dock are different commitments.)

Draw the loop and mark each handoff:

1. Customer use
2. Empty-crate storage
3. Collection
4. Depot processing
5. Return to available stock
6. Next delivery

Then write the organization responsible for each stage. If ownership is unclear, leave it blank. An unanswered "who does this?" may be the most important finding of the early research.

2. Give each assumption to the person who has the evidence

A coverage map shows which knowledge is missing. It is not a quota of one interview per category.

A supplier's view of customer demand is a lead, not customer evidence. A customer's guess about depot operations deserves the same caution. People can be accurate about their own work and mistaken about another organization's constraints.

The SBA's competitive-analysis guidance includes indirect alternatives and the influence of suppliers and customers. Use those relationships to look past businesses that happen to share your product's label. Read the SBA's market research guidance.

3. Ask about one recent handoff, not their whole business

Start with a recent delivery or collection the person can describe. Ask what arrived, what information came with it, what they did next, and what made the handoff complete.

• Customer side: "When the last delivery was unpacked, where did the packaging go, who handled it, and when did it leave?"
• Collection side: "What had to be ready before you could collect an item from that location?"
• Depot side: "What condition must incoming stock meet before it enters the next stage?"

These questions expose mismatches. The customer might assume the driver will pick up loose crates from several spots, while the route planner assumes one prepared collection point. Both accounts matter, and the gap belongs in the service design. A note saying "customer enthusiastic, planner cautious" loses it.

You don't need confidential operating records to ask about the general requirements of a handoff. Where a specification or rule matters, find the responsible public or authorized source and check it separately.

4. Choose the next interview by what blocks the next test

Start with the uncertainty that could most change the scope. If customers have no reason to change their current arrangement, optimizing the depot is premature. If there is a hard delivery constraint, customer enthusiasm can't fix it.

In the example, a first customer call might show that storage space is the real limit. The next useful interview is then a collection operator, to learn which pickup arrangements are feasible. A later customer call can test a concrete arrangement instead of a vague idea about reusable packaging.

Park interesting questions that don't affect the next decision. Write dependencies plainly: "Before testing this offer, establish who prepares the crates for collection." That makes the plan useful to the people building the service.

5. Keep the source on every claim

Tag each claim on the combined map as a firsthand event, general opinion, public document or untested assumption. Record the operating context and period, so a difference isn't mistaken for an error.

If two partners both say customers want the service, check whether they mean the same customers or the same industry report. One claim repeated through a network can look like independent confirmation.

Keep commercial interest visible without treating it as disqualifying. A supplier who benefits from your purchase may still give excellent technical context. Judge the support for the claim, and get a counter-view where the decision warrants it.

6. Finish with a readiness map

Mark each important handoff with one of three states: understood well enough for the next test, unresolved, or contradicted by current evidence. Add the next observation or conversation that would change its state.

The crate founder might be ready to test a limited collection arrangement with one kind of customer, while depot scale and broader demand stay unresolved. That is a bounded result. It is also stronger than calling the market validated because several stakeholders liked the idea.

A teammate should be able to trace the service from customer need through delivery and back, see who supports each assumption, and spot the missing link. Start by drawing your own loop, and leave the blanks blank.