What Counts as Real Demand Before Launch?

Positive feedback is a useful beginning. A customer understands the idea. The problem sounds familiar. The proposed solution feels relevant. Those reactions help a team learn the customer’s language and decide which parts of the concept deserve more attention.

They do not need to be dismissed simply because they are not a purchase. They do need to be interpreted for what they are: evidence of interest, not yet evidence of demand.

Demand becomes more credible when customers face a realistic choice and still move toward the offer. That may mean spending time, accepting a tradeoff, choosing it over an alternative, changing a habit, involving another decision-maker or exchanging money.

The research should become more demanding as the launch decision becomes more expensive. A few strong interviews may justify developing a concept. They should not, by themselves, justify a national launch, a large inventory commitment or a major sales investment.

Use an Evidence Ladder

A launch team can organize evidence into five levels. Each level answers a different question. The common mistake is using research from an early rung to support a decision that belongs much higher on the ladder.

The ladder is not a claim that every launch must use every method. It is a way to match the evidence to the decision. The right question is: What is the smallest credible study that reduces the most important uncertainty we face next?

A shoe launch through the ladder

Imagine launching a stylish walking shoe designed to provide all-day comfort without looking like a traditional running shoe.

1. Recognition: healthcare professionals describe foot fatigue during long shifts and the compromises they make with running shoes, clogs or insoles.
2. Engagement: they wear a prototype during real shifts, complete short usage diaries and return to discuss comfort, appearance and workplace requirements.
3. Choice: they choose among the new shoe and familiar alternatives at realistic prices, colors and feature combinations.
4. Purchase: qualified customers place a preorder, pay a refundable deposit or purchase through a limited market release.
5. Repeat: purchasers continue wearing the shoe, recommend it, buy another pair or choose the brand again when replacement time arrives.

Level 1: Learn Whether the Problem Is Real

Recognition research

Start with recent behavior, not reactions to the pitch

At this level, the job is to understand the customer’s world before asking them to evaluate the idea. The most useful methods are exploratory interviews, contextual observation and short diary studies.

How it is conducted

Recruit people who genuinely experience the situation you care about. Interview distinct segments separately. Ask participants to reconstruct the last time the problem occurred: what triggered it, what they did, what it cost, who else was involved and why they chose the current workaround. When possible, observe the situation or ask participants to record it as it happens.

What to look for

Repeated situations, meaningful consequences, existing spending, workarounds, dissatisfaction and recognizable triggers. A useful finding is not simply “people have this problem.” It is a clearer view of which customers experience it intensely enough to act.

What it cannot prove: Interviews can establish language, context and plausible need. They do not show how common the need is or whether people will choose the proposed product at a real price.

For an early qualitative study, eight to twelve well-matched interviews within a clearly defined segment will often reveal recurring patterns. That is a practical planning range, not a universal rule and not a statistically projectable sample. If several very different segments are being considered, recruit and analyze them separately rather than mixing everyone together.

Level 2: See Whether Customers Engage

Engagement research

Put something concrete in front of the customer

Once the problem and target segment are reasonably clear, move from conversation to experience. A concept test can evaluate a written or visual proposition. A prototype test shows how customers respond to an early version. A product-use test lets them use a physical or digital product in the setting where it would actually matter.

How it is conducted

Create a consistent concept that explains the customer problem, proposed benefit, important features and expected price or price range. In qualitative work, ask participants to interpret it in their own words and identify what is compelling, unclear or missing. In quantitative work, show concepts consistently—often one concept per respondent when direct comparison would bias the rating—and measure relevance, uniqueness, credibility and intended action.

For prototypes and samples

Give participants a realistic task or usage period. Define what they should try without coaching them toward a positive answer. Capture completion, time spent, repeat use, requests for help, abandonment and voluntary follow-up. Combine observed behavior with a short interview or diary so the team understands why the behavior occurred.

What it cannot prove: Engagement shows that an idea earns attention or performs in use. Free samples, prototypes and research incentives remove some of the friction customers will face in the market. Engagement is stronger than a compliment, but it is still not a purchase.

A consumer concept study may require a larger quantitative sample to compare segments or concepts reliably. A B2B study may use a smaller sample because the buyers are specialized and difficult to reach. In either case, sample size should follow the decision, the expected differences and the level of confidence required—not a standard number applied to every study.

Level 3: Introduce Alternatives, Price and Tradeoffs

Choice research

Make the new idea compete

Products rarely enter an empty market. Customers compare them with familiar brands, current vendors, internal workarounds, a cheaper option or doing nothing. Choice research becomes more credible when those alternatives are present.

Competitive or shelf-choice test

Place the concept or product in a realistic choice set with actual alternatives, prices, pack sizes, claims or service levels. Ask participants to choose—not just rate—then vary placement, messaging or price across controlled groups. For consumer products, this might resemble a shelf, shopping page or simulated store. For B2B, it may be a structured vendor comparison using realistic buying criteria.

Conjoint or discrete-choice research

Respondents complete a series of carefully designed choice tasks in which features, service levels, brands and prices change. Statistical modeling estimates the relative influence of each attribute and allows the team to simulate likely preference for alternative offers. Good design requires realistic attributes, believable levels and enough variation to estimate tradeoffs.

Gabor-Granger pricing

Show a specific offer at a series of prices and ask purchase likelihood at each price. The result can help estimate a stated demand curve and revenue implications. It works best when the offer is concrete and respondents understand the category. It remains stated intent, so it should be calibrated against behavior when possible.

Van Westendorp pricing

Ask when a price feels too cheap, inexpensive, expensive and too expensive. This is useful for understanding price expectations and an acceptable range, particularly early in development. It does not reveal a profit-maximizing price and should not be treated as direct evidence that customers will buy.

What it cannot prove: Well-designed choice research is much stronger than asking “Would you buy it?” It still takes place in a research setting. The resulting preference shares or purchase likelihood estimates should not automatically be called a sales forecast.

Level 4: Ask for a Real Commitment

Purchase research

Test whether interest survives real friction

Purchase evidence is strongest when the customer knows what is being offered, sees the actual price and gives up something meaningful. That may be money, budget, time, data, internal political capital or a place in an implementation schedule.

Consumer and CPG approaches

Use a limited release, geographic test, controlled e-commerce test, preorder or refundable deposit when the product and fulfillment plan are credible. Track the full funnel: qualified exposure, product-page engagement, add-to-cart, checkout, purchase, cancellations and returns. A landing-page click is engagement; a completed order is purchase evidence.

B2B approaches

Offer a paid pilot or a clearly defined commercial trial with scope, price, buyer and success measures. Track movement from user interest to decision-maker involvement, procurement, data access, implementation effort and signed agreement. A meeting request is useful, but it is not equivalent to budget approval.

What it cannot prove: A limited purchase test may overrepresent early adopters, unusually motivated buyers or customers attracted by a launch incentive. It demonstrates that some customers will buy under specific conditions; it does not yet establish repeatability at scale.

Tests that collect money or customer information should be honest about availability, timing, refunds and fulfillment. A fake purchase path may measure clicks, but it can damage trust and create legal or ethical problems. The strongest test is realistic without misleading the customer.

Level 5: Measure Whether Demand Continues

Repeat and retention research

Follow customers after the first yes

A launch can generate trial through novelty, promotion or sales effort. The more durable question is whether customers keep using, buy again, renew, expand or recommend the offer after experiencing it.

How it is conducted

Define cohorts based on when, where and why customers entered. Track usage, time to second purchase, reorder, churn, renewal, expansion, returns and support needs over a period that matches the natural buying cycle. Follow up with customers who continued and those who stopped so the behavioral data has an explanation.

What to compare

Compare target segments, acquisition channels, price points and use cases. Strong overall retention can hide a segment that is failing; weak averages can hide a small group with excellent product-market fit. The most valuable result often identifies who is succeeding and under what conditions.

What it cannot prove: Early retention does not guarantee that the business can acquire customers profitably or defend the market. It does show whether the product continues creating enough value for customers to stay.

Match the Method to the Question

Research becomes easier to design when the team starts with the uncertainty rather than a favorite method. A survey is not automatically the answer. Neither is an interview, conjoint study or pilot. Each is useful for a different decision.

Build a Research Sequence, Not a Single Study

One method rarely carries the entire launch decision. A practical sequence moves from inexpensive learning to increasingly realistic behavior. Each study should remove uncertainty before the company spends more.

1. Define the decision and the risk. Write down what the company is deciding, what must be true for the launch to work and which assumption would be most damaging if it were wrong.
2. Explore the customer’s current world. Use interviews, observation or diaries to understand the situation, language, consequences, alternatives and triggers. Use the findings to refine the segment and concept.
3. Measure and compare. Use quantitative research when the team needs to estimate how common a need is, compare segments or evaluate concepts consistently. Do not ask a qualitative sample to provide market percentages.
4. Introduce tradeoffs. Test the offer against realistic alternatives, features and prices through shelf choice, conjoint, pricing work or a structured vendor comparison.
5. Ask for behavior. Run a limited release, preorder, paid pilot or controlled sales test. Track the whole decision funnel, not the most flattering metric.
6. Follow the customer. Measure continued use, repeat purchase, retention, renewal, returns and advocacy. Learn why customers stayed and why others left.

Five Rules for More Credible Demand Research

Keep the test kind and honest

The goal is not to pressure customers into proving the idea. It is to create a realistic choice, make participation worthwhile and learn from the answer the customer is comfortable giving. If the test involves money, personal information or a promised product, be clear about what will happen next.

Read the Signal Without Overreacting

A test that produces limited choice or purchase is not a failed research project. It may show that the need is not urgent, the segment is too broad, the offer is unclear, the price is wrong, the timing or channel does not fit, a familiar alternative is stronger or the requested next step is too large.

Look for consistency across methods. Interviews may show that a problem is real. Concept testing may show which version communicates the value. Choice research may show whether the idea survives alternatives and price. A pilot or limited release may show whether customers act. Repeat behavior may show whether the value lasts.

Those methods do not need to produce identical numbers. They should begin to tell a coherent story about who will act, why they will act, what they will give up and what the business must do to help them succeed.