# Customer Development (Custdev): How to Test Hypotheses

- URL: https://vladimirnovozhilov.com/en/blog/customer-development-custdev-proverka-gipotez/
- Author: Владимир Новожилов
- Published: 2026-08-31
- Category: Market research

> Customer development (custdev) is a methodology for testing hypotheses about customers and markets through interviews and experiments before making large investments in a product. The method was created by Steve Blank and consists of four stages: customer discovery, customer validation, customer creation, and company building. The article covers the process, a custdev interview script, a preparation checklist, tables, case studies, and an analysis of common mistakes.

Most products fail not because they were built poorly, but because they built something nobody needed. **Customer development** (custdev) is a discipline that helps you avoid this trap: you test hypotheses about customers and the market before investing months into development. In this article, we'll cover what custdev is, what stages the process consists of, how to run a custdev interview, how hypothesis testing works in custdev, and which mistakes most often kill the results.

## What is customer development (custdev): definition and core idea

**Customer development is a methodology for systematically testing hypotheses about customers, their problems, willingness to pay, and sales channels through direct conversations and experiments.** In simpler terms, instead of making assumptions "in the office," you go out to real people, ask the right questions, and gather facts on which you build your product and business model.

The core idea of the method is captured in Steve Blank's phrase **get out of the building**. No market insight is born at a desk — it appears in conversations with people who already face the problem you want to solve.

It's important not to confuse custdev with product development. Product development is about *how* to build the product (sprints, releases, code). Customer development is about *what and for whom* to build, and *whether there's demand for it*. These two processes run in parallel: while the development team builds the solution, the customer development team checks whether it's needed at all.

![How effort is distributed across the stages of customer development](./images/customer-development-custdev-proverka-gipotez-stages.png)

## History of the method: how Steve Blank created customer development

The term was coined by entrepreneur and Stanford lecturer **Steve Blank** in his book **The Four Steps to the Epiphany** (2005). After going through eight startups, he noticed a pattern: teams pour all their energy into executing the product but rarely test basic assumptions about the customer. Later, together with **Bob Dorf**, he turned the method into a practical guide, "The Startup Owner's Manual."

Blank's ideas were further developed by his student **Eric Ries**, the author of the **Lean Startup** concept. It was Ries who connected customer development with iterative development and the "build — measure — learn" loop, making the approach mainstream.

## Customer development as part of Lean Startup: connection to the business model and Agile

Custdev isn't a standalone technique — it's part of the broader **lean startup** philosophy. The logic goes like this: a startup isn't a smaller copy of a big company, but a temporary organization searching for a repeatable and scalable **business model**. Until that model is found, any investment in scaling is a risk.

To describe and test the model, teams use Alexander Osterwalder's **Business Model Canvas** — a one-page diagram made up of nine building blocks: customer segments, value proposition, channels, customer relationships, revenue streams, key resources, activities, partners, and cost structure. Each block of the canvas is a set of **hypotheses**, and customer development is the way to test them.

The connection to Agile is simple: Agile is responsible for the speed of delivery (short iterations), while custdev is responsible for direction (what exactly to deliver). Quickly building something nobody needs is the worst thing you can do.

## The 4 stages of customer development: discovery, validation, creation, company building

The process is divided into four sequential stages. The first two are about *searching* for a business model; the last two are about *executing* and scaling it.

1. **Customer discovery.** You formulate hypotheses about the problem and segment and test them in interviews. The goal is to confirm that the problem is real and painful. No product is needed at this stage.
2. **Customer validation.** You check whether people are willing to pay for the solution. This is where a **minimum viable product (MVP)**, first sales, landing pages, and pre-orders come in. The outcome is confirmed **product-market fit** or a signal to pivot.
3. **Customer creation.** The model is validated — now you scale demand through marketing and sales, building a funnel and acquisition channels.
4. **Company building.** The startup turns into a stable organization: departments, processes, and hiring for scale appear. Search mode gives way to execution.

If a hypothesis isn't confirmed at the discovery or validation stage, you do a **pivot** — you change the segment, problem, or solution and go back a step. That's normal, not a failure.

### Comparison table of the 4 custdev stages

| Stage | Goal | Methods | Metrics / signals | Result |
|---|---|---|---|---|
| Customer discovery | Find a real problem and segment | In-depth interviews, observation, "getting out of the building" | Recurrence of the problem, emotions, facts about past behavior | Problem hypothesis confirmed |
| Customer validation | Check willingness to pay | MVP, landing page, pre-sales, letters of intent | Conversion rate, actual payments, early customer retention | Product-market fit found |
| Customer creation | Scale demand | Paid traffic, content, sales team | CAC, LTV, funnel growth rate | A working acquisition engine |
| Company building | Build a sustainable company | Hiring, processes, org structure | Unit economics, margin, churn | Scalable business |

## The customer development manifesto: Blank's 14 principles

The **customer development manifesto** is a set of reminder-style rules formulated by Steve Blank. Here's the essence in 14 points:

1. There are no facts inside the building — get out and talk to customers.
2. Combine customer development with agile development.
3. Failure is an integral part of the search for a business model.
4. Iterations and pivots are the norm, not the exception.
5. No business plan survives first contact with customers.
6. Design experiments and tests to validate hypotheses.
7. Agree on the type of market — it changes everything (new, existing, resegmented).
8. Startup metrics differ from those of a large company.
9. Fast decision-making, speed, and tact matter.
10. Everything comes down to the team's passion and resilience.
11. Job titles in a startup don't equal job titles in a corporation.
12. Conserve cash while searching for the model — spend on scale afterward.
13. Share and learn from insights gathered from customers.
14. Success in customer development starts with the whole team's shared belief in the method.

## How to run a custdev interview: script and rules for asking questions

**A custdev interview is a deep one-on-one conversation whose goal is to understand a person's real experience, not to sell them an idea.** The main rule: ask about past and present behavior, not hypothetical future intentions. People are bad at predicting what they'll buy, but they remember well what they've actually done.

![Hypothesis-screening funnel: only a handful of ideas make it through to confirmed status](./images/customer-development-custdev-proverka-gipotez-funnel.png)

### Custdev interview script: question template

**Warm-up (context):**
- "Tell me about what you do and what your typical workday looks like."
- "When was the last time you ran into [the problem]?"

**The problem (core of the interview):**
- "Walk me through how you dealt with this issue last time."
- "What was the most frustrating part of that process?"
- "How much time and money did it take?"
- "What did you try before that, and why did you stop?"

**Current solution and alternatives:**
- "What tool or method do you use now?"
- "If this tool didn't exist, what would you do?"

**Willingness to act:**
- "Have you looked for another solution? What exactly did you search for?"
- "Who in your company decides on purchases like this?"

**Closing:**
- "Who else should I talk to about this?"
- "Can I follow up with you once we have something to show?"

Notice that the script doesn't mention your product at all. At the customer development interview stage, you're studying the customer's world, not pitching a solution.

### Checklist for preparing a custdev interview

**Before the interview:**
- Formulate 1–3 hypotheses you're testing in this specific conversation.
- Define the profile of your interviewee and the criteria they must meet.
- Prepare an open-ended guide of 8–12 questions (not a survey).
- Get agreement to record the conversation.

**During the interview:**
- Talk 20% of the time, listen 80%.
- Dig deeper: "why?", "tell me more," "and then what?"
- Record facts and quotes verbatim — don't interpret on the fly.
- Hold your pauses — people often reveal the important part in silence.

**After the interview:**
- Within an hour, write down key insights and direct quotes.
- Note whether the hypothesis was confirmed, disproved, or needs clarification.
- Log the data into a shared table or CRM.

![After 10–12 interviews, new insights become increasingly rare](./images/customer-development-custdev-proverka-gipotez-saturation.png)

## Testing hypotheses in custdev: quantitative and qualitative data

**Hypothesis testing in custdev** relies on two types of data, and you need both.

**Qualitative data** answers the question "why": interviews, observations, open-ended answers. It gives you depth and reveals motivation, but it can't be statistically generalized.

**Quantitative data** answers the question "how much": landing page conversion rates, payment share, A/B tests, retention. It gives you scale and objectivity, but doesn't explain the reasons behind it.

The right sequence: first go qualitative (interviews at the customer discovery stage) to find patterns, then go quantitative (MVP and experiments at the customer validation stage) to confirm them. Relying on numbers alone without conversations is a common mistake — you end up optimizing something you don't actually understand.

### Mini hypothesis-tracking table template

| Hypothesis | How we test it | Success criterion | Result | Decision |
|---|---|---|---|---|
| Small businesses lose 5+ hours a week on manual reporting | 10 interviews | 7 out of 10 confirm the pain point | 8 out of 10 confirmed | Move forward |
| Willing to pay $20/month for automation | Landing page + pre-order | Conversion rate ≥ 3% | 1.2% | Pivot: lower the price / change segment |

This kind of table is convenient to keep in Google Sheets or a CRM, so insights aren't lost and are accessible to the whole team.

## Who should own custdev on the team: roles and responsibilities

Custdev is a team sport, but the proportions differ:

- **Founders and leadership.** In the early stages, founders themselves should run the first interviews — understanding the market can't be delegated.
- **Product and engineering.** Turn insights into hypotheses, MVPs, and backlog priorities.
- **Marketing and sales.** Own the customer creation stage: testing channels, messaging, and pricing.
- **Customer success and support.** A constant source of feedback from existing customers — "continuous custdev."

Rule of thumb: discovery-stage interviews shouldn't be outsourced. Whoever makes product decisions needs to hear the customer firsthand.

## Benefits of customer development: why companies should test hypotheses

- **Reduces risk.** You cut off nonviable ideas before making major investments.
- **Saves budget and time.** It's cheaper to change a hypothesis than to rebuild a finished product.
- **Gives you the customer's language.** Phrases from interviews become the foundation for positioning and website copy.
- **Speeds up product-market fit.** You move toward demand based on data, not intuition.
- **Builds a culture of fact-based decisions.** "I think" debates get replaced by testable hypotheses.

## Risks and pitfalls of custdev: common mistakes

**Mistakes in the process itself:**
- **Attachment to the idea.** You only hear what confirms your hypothesis.
- **Fear of pivoting.** The data says "change," but the team stubbornly keeps building the same thing.
- **Overreliance on quantitative data.** Numbers without explanation lead to wrong conclusions.
- **Overgeneralization.** One vivid piece of feedback gets treated as the opinion of the entire market.

**Mistakes during custdev interviews:**
- **Leading questions.** "Wouldn't it be more convenient this way?" — people agree out of politeness.
- **Hypotheticals instead of facts.** "Would you buy this?" instead of "What did you buy last time, and why?"
- **Selling instead of researching.** You pitch the product and stop listening.
- **Made-up answers.** The interviewee fantasizes about the future, but what you need are facts about past behavior.
- **Small sample or a single segment.** Three interviews with friends don't validate a hypothesis.

## Customer development examples: case studies from Dropbox, Airbnb, Zappos

**Dropbox.** Before building the complex file-syncing technology, Drew Houston released a short demo video showing how the product *would* work. Overnight, the waiting list grew from 5,000 to 75,000 people. This is a classic MVP experiment at the customer validation stage: demand was confirmed before full-scale development.

**Airbnb.** The founders literally followed the "get out of the building" principle: they visited hosts in New York, personally photographed apartments, and talked to people. These interviews revealed that photo quality directly affected bookings — an insight impossible to get from analytics alone.

**Zappos.** Founder Nick Swinmurn didn't build a warehouse or a logistics operation. He photographed shoes in regular stores, posted them online, and whenever an order came in, he'd go buy the pair himself and ship it. This let him validate the core hypothesis — whether people were willing to buy shoes online — at minimal cost. It turned out they were.

The common thread across all three cases: first validate demand cheaply, then scale.

## Custdev and product metrics: connecting it to churn and unit economics

Customer development doesn't end at launch. At the customer creation and company building stages, custdev merges with metrics and unit economics.

- **Churn** is a direct signal for a new round of interviews. Talk to customers who left — their answers will show you where the product falls short far more precisely than any analytics dashboard.
- **Retention** confirms or disproves the product-market fit you found: if people don't come back, the "fit" was an illusion.
- **LTV and CAC.** Testing channels and pricing at the customer creation stage is also hypothesis testing — just quantitative. Unit economics need to work before you scale.

According to industry benchmarks, companies that build continuous custdev into their work with existing customers manage churn noticeably more effectively — because they respond to causes, not symptoms.

## Tools and templates for custdev: CRM, spreadsheets, checklists

You don't need complex tools — discipline matters more. The basic toolkit:

- **Google Sheets / Notion** — a hypothesis registry and interview log (see the mini template above).
- **CRM** — storing respondent contacts, statuses, and notes so insights don't get lost and stay accessible to the team.
- **A video-calling tool with recording and transcription** — to capture quotes verbatim.
- **A whiteboard (Miro/FigJam)** — for affinity mapping: grouping insights into patterns.
- **Business Model Canvas** — as a map of hypotheses you're progressively closing out.

Automation here helps not with the interviews themselves, but with processing them: transcripts, tags, consolidating data into a table. That frees up time for what matters most — the conversations.

## Custdev methodology: a short implementation algorithm

For the custdev methodology to actually work, rather than stay theoretical, follow these steps:

1. Map out the business model on a canvas and identify the riskiest hypotheses.
2. Start with customer discovery: 8–15 interviews per segment, focused solely on the problem.
3. Consolidate the insights and decide whether to confirm demand, change the hypothesis (pivot), or kill it.
4. Move to customer validation: build an MVP or landing page, test willingness to pay.
5. Once you've found product-market fit, scale demand (customer creation) and build the company (company building).
6. Don't stop there: turn customer conversations into an ongoing process.

Customer development isn't a one-time study — it's a way of thinking: every product decision rests on a tested hypothesis, not a guess. That's exactly how custdev saves resources and brings the product to the market, rather than trying to drag the market to the product.

## FAQ on customer development and custdev

The main questions are covered in a separate section above — they address queries like "what is custdev," "how many interviews are needed," "which questions to avoid," and "can you do custdev without an MVP." Use them as a quick cheat sheet before launching your own hypothesis-testing cycle.

## FAQ

### What is custdev in simple terms?

Custdev (customer development) is a methodology for testing hypotheses about customers through conversations and experiments. Instead of building a product 'in a vacuum' for years, you go out and talk to real people, find out their problems and willingness to pay, and only turn validated hypotheses into features. The method was created by Steve Blank in his book The Four Steps to the Epiphany.

### How is customer development different from market research?

Classic market research more often answers the question 'what is the market and how much money is in it' through surveys and statistics. Customer development is more focused: it's about testing specific hypotheses about a problem, segment, and solution through in-depth interviews and MVP experiments. Custdev is iterative and tied to a decision — continue, pivot, or kill the hypothesis.

### How many interviews do you need to test a hypothesis?

A practical rule of thumb is 8 to 15 interviews per segment. Usually after 5–7 conversations you start hearing recurring patterns, and by 10–12 you reach 'saturation,' when new insights are rare. If answers are still all over the place, the segment is too broad — narrow it down and gather more interviews.

### What questions should you avoid asking in a custdev interview?

Avoid leading and hypothetical questions like 'Wouldn't this feature be useful to you?' or 'Would you buy the product for $10?' People agree out of politeness, and you end up with made-up answers instead of facts. Ask about past experience: how they solved the problem last time, how much time and money they spent, and what they tried before.

### Can you do custdev without an MVP?

Yes. The first stage — customer discovery — doesn't require a product at all: you're exploring the problem through interviews. An MVP (minimum viable product) is needed later, at the customer validation stage, to check whether people will actually pay for the solution. Often a landing page, a prototype, or a manual service plays the role of an MVP, as in the Zappos case.