Validate a New Offering with Voice of Customer Research

To validate your assumptions, interview a purposeful mix of current customers and prospects, including both buyers and users whose needs may differ. Ask how they experience the problem, what prompts them to act, how they currently solve it, and what resources they commit to finding a solution.

As you collect this information, use AI to synthesize the conversations, compare findings across interviews, and identify recurring needs or evidence that contradicts your hypotheses. We call this voice of customer research. It is a structured process for gathering and analyzing direct customer input about the problem, the alternatives used, and the value of a better solution. This evidence helps you determine whether your first client’s need represents a repeatable market opportunity or an isolated case, ultimately clarifying whether further investment is justified.

Why doesn’t one client prove demand?

At least once a week I get on a call with someone new, and they tell me all about their latest “product.” They built it for their best client and were sure it would have broader appeal but, for some reason, it’s just not selling. Unfortunately, this is particularly common among professional services companies.

Responding to individual client requests is how you win work in a services business. One client’s willingness to pay can justify developing a solution tailored to their needs. That makes a request from your best client a credible reason to build, but it can also encourage you to assume that other buyers will follow. Before investing in a repeatable offering, you need evidence that customers beyond that relationship share the problem and will pay for the same solution.

We refer to this risk as the Problem of One but not Many. One training company experienced this when a client requested virtual delivery of a nine-class program. Seeing it as a great opportunity to grow revenue with a new offering, the firm invested six figures into an online platform without verifying broader demand, only to discover that the solution appealed only to that original client and now they were committed to a big, annualized tech expense.

Nonetheless, first buyers remain a source of valuable evidence. You can learn from their purchase, observe the performance of their offering, and identify which aspects of their experience deliver the most value. Leverage the insights, but be careful not to let the specific requirements of that one buyer define your entire offering before other customers have also confirmed the need.

A Journal of Marketing meta-analysis supports this distinction. It examined 123 relationships reported across 35 empirical studies. It found that customer participation during ideation and launch was positively associated with financial performance and faster time to market. During development, however, customer participation was negatively associated with both measures. This indicates that services firms should prioritize customer input during the ideation and launch phases to validate the problem, but not allow a single client’s specific preferences to steer the development process.

What must your voice of customer research establish before you build?

Before you begin interviews, convert your assumptions into hypotheses that the evidence can support or contradict. Without this framework, your team might gather interesting anecdotes without ever resolving the critical questions that should determine if the offering merits investment.

To avoid this, document your beliefs in three specific areas:

1. Describe the urgent and expensive customer problem. Identify the main problem, its root cause, why customers need to solve it within the current calendar year, and whether they would allocate budget or dedicated resources to solving it.

2. Identify the customer segments experiencing the problem. Name the target buyers and users, including any meaningful differences among segments or personas.

3. Understand how customers currently solve the problem. Record the solutions they use today, whether they are satisfied with them, and where those alternatives fall short.

These hypotheses keep your research focused. The next step is to build open-ended questions around them so that customers can describe their experience without being led toward your preferred answer. When a participant describes a problem, dig deeper by asking why it occurs and what the consequences will be if it remains unresolved. This progression moves the conversation from a visible symptom toward its underlying cause, helping you determine whether customers across your target segments share the same problem.

Even if customers confirm that the problem is urgent and expensive, the market might still be too narrow to support your business goals. For example, one copywriting services firm invested over $100,000 to transform an internal training program into an external product, only to discover that the target market was limited to a very small group of companies with enough in-house copywriters to require it. In the end, there just wasn’t enough market demand to justify the cost to build.

You can avoid the same costly mistake by estimating the total addressable market before committing to a larger investment. Multiply the number of potential customers by a realistic average price to estimate the theoretical revenue opportunity. Then create a preliminary business case using the share of that market you could realistically reach, the expected revenue, and the cost to build and deliver the offering. Compare those estimates with your firm’s growth goals.

How many customer interviews are enough?

Before determining the number of interviews to conduct, start by segmenting your potential buyers. Divide them based on characteristics that might influence their buying behavior, such as revenue, the maturity of their organization, or their title.

Once you have established these segments, plan for five to six interviews within each relevant segment. Compare responses within each group before looking for patterns across groups, since buyers at different company sizes or levels of maturity may have different needs.

The composition of your sample is far more important than the total count. If your goal is to attract a new market segment, interviewing only your existing customers leaves a critical assumption untested. Similarly, speaking only with VPs of Talent at large enterprises when the actual buyer is the CHRO at mid-market firms will send you in the wrong direction.

Frequently cited studies like the Field Methods study suggest that basic themes often emerge within six interviews and saturation occurs within 12. We agree. When you can predict what you are about to hear, stop, regroup, and pivot if needed. Consider what new hypotheses have formed. Are there other questions you should be asking?

To ensure your research stays on track, review your findings after every two or three interviews. If the evidence challenges a core assumption, update your interview guide before proceeding. Finally, remember that consistency is key.

For more help planning and conducting these conversations, read our five tips for customer interviews.

How do you test demand beyond what customers say?

Interviews reveal if customers recognize a problem, how they currently solve it, and what outcomes they value. However, expressing interest in a solution is not the same as committing to buy one.

Since interviews only confirm the problem, you need a different test to confirm demand. Once your research supports your core hypotheses, launch a minimum viable product, or MVP, to test your offering without committing to a full build. The most effective MVP asks customers to pay for the solution under realistic conditions.

1. Use sell-then-build to secure commitment. Present your concept to qualified buyers and ask them to make a genuine commitment before you finish development. This might mean a paid pilot for a premium offering or a preorder for a lower-priced one. As our sell-then-build guidance suggests, secure multiple buyers before you begin the build to prevent the first client’s specific requests from turning the offering into bespoke work.

2. Use a Wizard of Oz MVP to validate the experience. Manually perform the tasks that technology would eventually automate while delivering the intended customer experience. This approach helps you determine which aspects of the offering create the most value before you invest in the technology. Our Wizard of Oz MVP guide details why you must define your success and failure criteria before starting this test.

3. Use a concierge MVP to refine delivery. Inform a small group of customers that experts will deliver the offering directly and work closely with them through the process. This transparency allows you to ask direct questions while learning which manual steps can be standardized for a repeatable product.

How should AI support voice of customer research?

AI can accelerate your research if you give it a clear, limited role. Use it to synthesize sales conversations, interview notes, and delivery feedback, then compare the findings with the hypotheses your team has already established. However, while these capabilities are powerful, they cannot replace direct interaction with your customers.

A synthetic panel, which simulates people in a target segment, may offer superficial similarities to human data, but it lacks the necessary context. A study published in Political Analysis found that synthetic survey responses contained less variation than human responses and changed when researchers made relatively small adjustments to the prompt. Although this study analyzed political surveys rather than B2B purchasing, its findings support using synthetic feedback only to generate or challenge hypotheses, not confirm demand. You must treat AI as an organizational tool. Let the technology surface contradictions and organize your evidence, then use your own judgment to verify its conclusions through real conversations and market tests.

What evidence is strong enough to support further investment?

Even with a well-organized research process, no single interview or AI-generated summary can make the investment decision for you. To move forward with confidence, you must compare the combined weight of your real-world evidence against the hypotheses you established at the beginning.

Before approving a larger build, answer five questions:

1. Do customers across the relevant segments describe the same underlying problem, and does it pass the urgent and expensive test?

2. Is the target market large enough to support the intended investment and commercial model?

3. Have multiple qualified buyers made meaningful commitments under realistic conditions?

4. Has a limited test shown which parts of the offering can be repeated and where expert involvement remains necessary?

5. Do the combined results meet the success and failure criteria established before the test began?

Define these benchmarks before anyone sees the results. For each important hypothesis, state what evidence would cause the team to revise the target segment, reduce the scope, change the delivery model, delay the investment, or stop. Without predefined criteria, the team cannot distinguish a weak result from one that supports continued investment. A decision to stop does not invalidate the research; instead, it prevents the firm from compounding an unsupported assumption with a larger, riskier investment.

Which validation mistakes can distort the evidence?

To ensure your research provides an accurate reflection of the market, avoid these four common pitfalls that often distort evidence:

1. Do not let the first buyer dictate the scope of your offering. While their specific requests may improve your initial engagement, they can inadvertently turn a scalable product into a bespoke service that is difficult to repeat for others.

2. Do not lead with reactions to your solution. Instead, ask problem-led questions that reveal what customers already value and fund, which establishes an unbiased baseline before your concept influences their responses.

3. Do not rely exclusively on your current clients for research. Since a new offering is designed to attract different types of buyers, you must include prospects in your sample to validate that the need exists beyond your existing relationships.

4. Do not treat AI output as definitive market evidence. While AI is excellent at synthesizing data to clarify your research plan, it cannot replicate the real-world commitment of a customer choosing to buy.

By avoiding these traps, you ensure that your investment decisions are based on genuine market demand rather than bias. Staying disciplined throughout your validation process creates the necessary foundation to build a solution that customers are truly eager to purchase.

A single client can open the door to a valuable new offering, but the next step is determining whether a scalable market exists on the other side. By balancing AI-assisted synthesis with disciplined human-led research, you can build with confidence rather than speculation. If you want to pressure-test your validation plan before committing to a larger investment, start a conversation with us.