Customer development is a method for testing your business assumptions with real customers before you spend heavily on building, selling, or scaling. Steve Blank created it for startups. Any team launching something new can use it.
The stakes are high. CB Insights found that 43% of the venture-backed startups it analyzed that shut down since 2023 cited poor product-market fit. Product-market fit means enough customers want what you built.
In this blog, we’ll cover the four stages, the interview questions that work, the signals that show progress, and real examples.
What is customer development?
Customer development is a four-step, hypothesis-driven framework for validating a business model through direct conversations with potential customers. Steve Blank designed it so founders can replace guesses about customers with evidence before they scale. Practitioners often shorten it to CustDev.
A hypothesis is a testable guess. For example: “Dental offices with fewer than five staff lose hours each week to phone rescheduling.” You write the guess down, then try to prove it wrong by talking to the people it describes.
Blank’s guiding idea is that there are no facts inside the building. Answers live with customers, so founders have to “get out of the building” to find them.
According to the researchers, the framework checks three things. A real customer need exists. Your product meets it. Your organization can scale to serve the demand.
Where did customer development come from?
Customer development grew out of Blank’s experience as a serial founder and became a pillar of the Lean Startup movement. Three milestones shaped how teams use it today.
- 2005: Blank published The Four Steps to the Epiphany, which laid out the original four-step model, according to the Agile Alliance.
- 2011: Eric Ries popularized the approach in The Lean Startup, which pairs customer development with agile engineering.
- 2013: Blank explained the idea to a wider business audience in the Harvard Business Review article Why the Lean Start-Up Changes Everything.
Lean Startup combines three parts. A business model canvas frames your hypotheses. Customer development tests them. Agile engineering builds the product in small steps.
Why does customer development matter?
It reduces the biggest early risk: building something nobody wants. The same CB Insights report analyzed 431 failed venture-backed companies. Many cited several causes, so market fit problems often sit beside cash and timing problems.
Blank describes the alternative in HBR. Lean startups test their hypotheses with early customer feedback instead of following a long business plan. That shift lowers the odds that you build a product no one will pay for.
Three practical benefits follow. Decisions rest on what customers did and said, not on opinions in a meeting. Conversations cost far less than engineering time. And you gain evidence to show teammates and investors when you ask for resources.
How is customer development different from similar terms?
Several terms get mixed up with customer development. They overlap, but each answers a different question. Traditional research, including the low-cost options in this guide to market research for startups, usually measures an existing market. Customer development tests your specific assumptions inside it.
| Term | What it is | Core question |
|---|---|---|
| Customer development | A four-step framework for testing a business model with customers | Does a scalable business exist here? |
| Lean Startup | A broader method combining the business model canvas, customer development, and agile engineering | How do we build, measure, and learn fast? |
| Customer discovery | Step one of customer development | Does this problem exist, and who has it? |
| Customer validation | Step two of customer development | Will customers pay for our solution? |
| Market research | A broad study of market size, competitors, and preferences | How big is the market and what do buyers want? |
| Agile development | An iterative way to build the product | How do we build the solution incrementally? |
What are the four stages of the customer development process?
The customer development process has four stages. Stages one and two search for a repeatable business model. Stages three and four execute it.
- Customer discovery: Confirm the problem is real and find who has it.
- Customer validation: Prove customers will pay for your solution.
- Customer creation: Build demand and grow the customer base.
- Company building: Turn the startup into an organization that runs the model.
Stage 1: Customer discovery
Customer discovery tests whether the problem is real and whether your idea fits it. You test two hypotheses. The problem hypothesis says who has the problem and how much it hurts. The product hypothesis says your solution would fix it.
Rank your hypotheses by risk and test the one that could sink the business first. Common discovery methods include:
- Problem interviews with people in your target segment
- Observation of how they handle the task today
- Support tickets, forums, and reviews that show repeated complaints
- Short surveys that show which pain points appear most often
Our guide to customer pain points helps you sort which problems deserve attention.
Stage 2: Customer validation
Validation tests whether customers will pay for your solution. Compliments do not count. Commitments do.
Strong commitments include pre-orders, paid pilots, and a signed letter of intent (LOI), a short document stating that a buyer plans to purchase if you deliver. Many teams sell a minimum viable product (MVP), a basic version with just enough features to test demand. If few buyers commit, return to discovery. A viability study adds a structured check on cost and demand.
Stage 3: Customer creation
Creation starts once you have a repeatable way to win customers. Your spending shifts from learning to acquisition. Blank’s model treats the market type as a key choice:
- Existing market: Buyers already purchase something similar, so you compete on a clear difference.
- New market: Buyers do not yet know they need the product, so you must educate them.
- Resegmented market: you target a niche or offer a lower-cost version of an existing product.
Stage 4: Company building
Company building turns a team searching for a model into an organization that executes one. You add departments such as sales, marketing, and support. Processes replace founder improvisation.
The listening does not stop. Product, sales, and support teams keep feeding customer evidence back into decisions. Without that loop, the company slowly drifts away from the customers it set out to serve.
How do you know when to move from discovery to validation?
Move on when the evidence repeats, not when you hit an interview count. Counts still help as a guide. Blank’s classes ask teams for 100 customer conversations in 10 weeks, which shows how much listening serious discovery takes. For your own project, 10 to 15 interviews per segment is a workable start. Keep going until answers stop surprising you.
Use four criteria to decide:
- Pattern: Most people in one segment describe the same problem without prompting.
- Pain: they Spend time or money on workarounds today.
- Segment: You can name who they are and where to reach them.
- Pull: Several agree to a follow-up, a pilot, or an introduction.
If you miss these bars after about 20 conversations, change the hypothesis. That change is called a pivot, a deliberate shift in direction based on evidence. It is part of the method, not a failure.
Step-by-step guide to customer discovery interviews
A discovery interview is a structured conversation where you learn how someone handles a problem today. Plan 30 to 45 minutes and record with permission. For general technique, see this guide to in-depth interviews.
- Write your top three hypotheses in one sentence each.
- Recruit people who match your target segment, not friends or family.
- Run problem interviews first. Ask about past behavior and hold your idea back.
- Listen for emotion, workarounds, and money already spent.
- Write up notes within an hour and tag recurring themes.
- After every five interviews, update your hypotheses.
- Run solution interviews only after the problem pattern holds.
Which interview questions work best?
Rob Fitzpatrick’s The Mom Test offers three rules: talk about their life, ask about specific past events, and listen more than you talk. Opinions about your idea tend to be polite and useless. Wording also sways answers, as this guide to leading questions shows.
| Weak question | Stronger question |
|---|---|
| Would you use an app that does this? | Tell me about the last time you handled this task. |
| Do you think this is a good idea? | What have you already tried to fix this? |
| How much would you pay for it? | What do you spend on this problem today? |
| Do you always struggle with this? | When did this last cause you trouble? |
How do you measure customer development progress?
Measure behavior, not opinions. Six signals show whether your hypotheses hold. Set the pass mark before the interviews start, such as “at least 6 of 10 name the problem unprompted.” A bar written in advance stops you from explaining away weak results.
- Problem frequency.
The share of interviewees who raise the problem unprompted, tracked per segment.
- Problem intensity.
How much time or money they already spend on workarounds.
- Commitments.
Pre-orders, paid pilots, and signed LOIs per 10 conversations. Compliments score zero.
- Customer acquisition cost (CAC).
Total sales and marketing spend divided by new customers won. Compare it with the revenue each customer brings in.
- Retention.
The share of customers who stay. A high churn rate, the pace at which customers leave, signals weak fit.
- Product-market fit survey.
A survey that asks users how they would feel without your product. See this product-market fit survey guide for question design.
Customer development examples that show each stage
Three examples show how the stages play out. Dropbox is a real case. The other two are illustrative scenarios built from common patterns, not named companies.
| Example | Stage in focus | Key lesson |
|---|---|---|
| Dropbox | Discovery and validation | Test demand cheaply before scaling spend |
| Scheduling software startup (illustrative) | Discovery | Interviews can change the target segment |
| Subscription box retailer (illustrative) | Validation | Pre-order tests reveal real pricing limits |
Example 1: Dropbox
Dropbox first tried paid search ads. According to Launch Tomorrow’s breakdown, acquisition cost was at least $233 per customer for a $99 product. Ad spend totaled about $3,000, and the channel failed.
Next, the founders built a demo video and sent Hacker News traffic to a landing page. That page collected around 5,000 interested prospects. Demand got tested before more building, at a fraction of the earlier cost.
Example 2: A scheduling tool (illustrative)
Founders assume solo salon owners will pay for booking software. After a dozen interviews, the picture changes.
- Assumption: Solo owners lose bookings and want a paid tool.
- What interviews showed: Solo owners happily use free apps, but multi-chair salons lose hours each week to no-shows.
- Change: The team pivots to multi-chair salons and tests a paid no-show reminder feature.
Example 3: A subscription box retailer (illustrative)
An established US retailer wants to launch a subscription box. Instead of building one, the team runs a pre-order page with three price points and emails it to existing customers. The middle price draws the most pre-orders, so the team drops the premium tier and launches with two.
The test cost a few days of work. A full build would have cost months. Validation protected the budget.
What are the most common customer development mistakes?
Most failures trace back to a handful of habits. Each one weakens the evidence you collect.
- Interviewing friends and family.
They protect your feelings instead of describing their needs.
- Pitching instead of asking.
Once you describe your idea, answers bend toward it.
- Treating compliments as validation.
“That sounds great” costs nothing to say.
- Stopping too early.
Three conversations show anecdotes, not patterns.
- Skipping validation.
Interest in an idea is not willingness to pay.
- Talking only to happy customers.
Include people who churned or never bought.
- Treating it as a one-time phase.
Markets shift, so the learning has to continue.
When is customer development not the right fit?
Customer development works best when uncertainty about customers is high. It has limits. Customers can describe problems well, but they struggle to imagine inventions they have never seen. For breakthrough products, combine interviews with prototypes people can react to.
Enterprise sales add a second limit. A buying committee includes users, budget holders, and gatekeepers, and each sees value differently. Interview all of them, not only your friendliest contact.
Regulated fields add a third. In US healthcare, the Health Insurance Portability and Accountability Act (HIPAA) limits how you can recruit and record patients. Plan consent and privacy steps before you schedule the first call.
How can you scale discovery findings with QuestionPro?
After interviews reveal a pattern, you need to confirm it across more people. Surveys do that. The harder part is learning why people answered the way they did, and open-ended boxes often return one-line replies.
Moderated (Chat) is a QuestionPro AI question type that asks follow-up questions based on each respondent’s answer. You set key learning objectives, a tone, and a maximum number of follow-ups. It is currently in beta, and each chat uses one QuestionPro AI credit. To add it:
- Open the survey editor and choose Add Question, then Advanced Questions, then Moderated AI, then Chat.
- Pick Follow-up mode to probe an earlier answer, or Independent mode to open a fresh topic.
- Enter your learning objectives and the maximum number of follow-ups.
Pair it with a validation survey on pricing or purchase intent, and you get both the numbers and the reasons. The wider QuestionPro Market Research Software also covers respondent sourcing and a place to store past findings.
Why customer development never really ends
Customer development is less a project than a habit. The four stages give you a path, but the real value comes from staying curious about customers after the launch. Markets shift, competitors copy, and yesterday’s validated problem can fade.
Treat every new feature, segment, or price change as a fresh hypothesis. Teams that keep asking stay close to what buyers want. Three steps get you started:
- Write one hypothesis you could be wrong about.
- Book five conversations with people who fit your segment.
- Set your pass or fail bar before the first call.
Frequently Asked Questions (FAQs)
No. Established companies use it when launching products, entering new segments, or testing major features. The Agile Alliance notes that it also applies to internal product work, wherever customer needs are uncertain and a wrong guess would be costly.
Start with LinkedIn, industry Slack groups, trade associations, and your own customer lists. For faster reach, a research panel can screen people by job role and location. Offer a fair incentive, such as a gift card, and confirm each person matches your segment.
No. Surveys confirm patterns across many people, but they cannot uncover problems you did not think to ask about. Use interviews to find the right questions, then run a survey to size how common each answer is.
It depends on your market. Blank’s classes target 100 conversations in 10 weeks, so a focused discovery round often takes two to three months part-time. B2B markets with long sales cycles usually need longer, and validation adds more time.
Treat it as useful data. Identify which part failed: the problem, the segment, or the solution. Change one variable, such as the target customer, and test again. Several rounds of small pivots are normal before a model holds up.
