Sign inGet started
Skills / Entrepreneurship

Customer Development

No business plan survives first contact with customers.

Customer Development is Steve Blank's four-step methodology for discovering and validating the market for a new product before wasting resources building the wrong thing. Designed for early-stage founders, intrapreneurs, and product managers, it replaces assumption-driven roadmaps with a structured, hypothesis-testing process grounded in direct customer conversations. Unlike generic 'talk to customers' advice, it provides specific frameworks for what to test, how to interpret what you hear, and when to pivot versus press forward.

By Steve Blank · Free
Specimen 01 · Live diagnosisCustomer Development
Input

“We're building a project management tool for architecture firms. I think the customer is the project manager or principal at small firms, 5–20…”

Diagnosis
Let's start where Steve Blank would: not with your solution, but with your hypotheses.
Full transcript ↓
Calibrated referenceagent-skills.ai
The gap

Stop building what nobody wants — get out of the building and find out first

Customer Development runs four sequential steps in parallel with product development: (1) Customer Discovery — founders 'get out of the building' (GOOB) to personally test problem, solution, and business model hypotheses through structured interviews, deliberately NOT selling; (2) Customer Validation — the team attempts to sell to a small group of 'earlyvangelists' to prove a repeatable, scalable sales process exists; (3) Customer Creation — proven demand is scaled into the sales channel, with strategy differentiated by market type (new, existing, re-segmented, or clone); (4) Company Building — the startup formalizes into a company, shifting from learning mode to execution mode. Each step produces specific artifacts: a tested Business Model Canvas, prioritized interview findings, an earlyvangelist profile, a validated sales script, and a demand-creation playbook. Pivot-or-proceed decision gates at the end of Discovery and Validation prevent teams from scaling a broken model.

The problem

Founders default to 'product development' mode — writing code, hiring, and spending — based on assumptions they've never tested with real customers. By the time a product launches, months of work are already locked into decisions that real customers would have invalidated on day one. The result is the most common startup failure pattern: a technically functional product that the market doesn't want.

The solution

Run your business model assumptions through a structured discovery process before committing to them. Come away with an interview script, a prioritized hypothesis list, and a clear pivot-or-proceed framework — so you spend the next month learning what's real, not building what isn't.

You bring
  • Your current startup or product idea with its core assumptions
  • Any existing customer conversations, user research, or market observations
  • A sense of which step of Customer Development you're currently in (or stuck on)
  • Willingness to have your hypotheses stress-tested rather than validated by default
You get
  • A prioritized list of business model hypotheses mapped to the Business Model Canvas, ranked by risk
  • A ready-to-use Customer Discovery interview script that tests problems without pitching solutions
  • An earlyvangelist profile defining the specific early adopter most likely to buy your unfinished product
  • A structured pivot-or-proceed decision based on your interview findings, with explicit criteria for each path
In action

Watch the methodology work.

Three specimens from a single real session: the same situation, unaided and calibrated, the full transcript, and the skill answering live in the channel where the work happens.

Fig.01The same situation, two ways
Specimens A · B
Specimen AVague
Unaided

“You've been coding for three months on a project management tool for architecture firms, based on your own experience at one firm. You have strong conviction about the problem but no external customer interviews. Your roadmap is built on 15 assumptions you've never tested — who the buyer is, whether they'll pay, whether they've already solved it with Procore or Newforma. You have four customer calls on the calendar and you're planning to demo the prototype.”

×Vague — nothing concrete to act on.
Specimen BCalibrated
With Customer Development

“Before those four calls, you've replaced the planned demo with a structured learning conversation. Each call surfaces whether the submittal/RFI chaos is widespread, who actually owns the pain, and whether firms are actively seeking a solution. After 12 interviews, you discover that the real decision-maker is the firm owner, not the PM, and that half the firms you interviewed already use Newforma — but hate it. That's a re-segmentation opportunity, not a new market. You pivot your positioning and earlyvangelist profile before writing another line of code.”

Specific, grounded, and ends with one clear next step.
Fig.02A real session, unedited
In-app · assistant
Customer DevelopmentClaude
Customer Development
Let's start where Steve Blank would: not with your solution, but with your hypotheses. Before we build anything, I want to map out the assumptions buried in your business model — because each one is a potential way to waste six months. What's the startup or product idea you're working on, and who do you think the customer is?
We're building a project management tool for architecture firms. I think the customer is the project manager or principal at small firms, 5–20 people. We've been coding for about 3 months.
Got it. Three months in is actually a good time to do this — you still have room to pivot without throwing away everything. Let me map out the riskiest hypotheses you're currently betting on. You've assumed: (1) architecture PMs have a project management problem that isn't solved by existing tools like Asana or Procore; (2) the pain is acute enough that they'd pay for something purpose-built; (3) the decision-maker is the PM or principal, not the firm owner or IT; (4) small firms (5–20 people) are actually a reachable, distinct segment. Which of these do you have real customer evidence for — meaning direct conversations, not assumptions from your own experience?
Honestly, mostly assumption. I used to work at an architecture firm so I experienced the pain firsthand, but I haven't done formal interviews yet. I figured I knew the problem.
That's the most common founder trap — and Blank calls it out explicitly. 'I am the customer' is a hypothesis, not validation. Your experience tells you the problem exists for you, at one firm, at one point in time. It doesn't tell you whether the pain is widespread, whether firms are actively seeking a solution, or whether they have budget and authority to buy. Here's what we need to do: before your next line of code, you need 10–15 Customer Discovery interviews. Not demos, not pitches — conversations where you're learning, not selling. Let me build you an interview script. First question: what's the specific moment of failure you experienced? What was happening when you thought 'there has to be a better way'?
It was tracking submittals and RFIs — there are hundreds of document exchanges between the architect, contractor, and client, and we were managing it all in email and spreadsheets. Things got missed, clients blamed us, it created liability.
That's a strong problem signal — liability exposure is a visceral pain. Now let's build your interview script around that. Your first probe isn't 'do you have this problem?' — that's leading. It's: 'Walk me through how your firm currently handles submittals and RFIs from the moment a request comes in to when it's resolved.' You're listening for: (1) Do they describe the chaos you experienced, or have they solved it? (2) What have they already tried — Procore, Newforma, manual tracking, nothing? (3) How painful is it — do they describe client blowback, liability claims, late nights? Anyone who's experienced all three and is still looking for a solution is a potential earlyvangelist. Want me to build out the full 12-question script, or should we first finish mapping your riskiest hypotheses on the Business Model Canvas?
Let's do the full interview script — I have calls set up with four firms next week.
Four calls is a start — Blank's target is 10–15 before you draw conclusions, so plan to keep scheduling. Here's your discovery script. Start every call with this framing: 'I'm doing research on how architecture firms manage project documentation and communication. I'm not selling anything — I want to understand how your firm works today.' Then move through these questions in order: [1] 'Tell me about a recent project. What does your documentation workflow look like from kickoff to closeout?' [2] 'Where do submittals and RFIs live in that process — who owns tracking them?' [3] 'Walk me through the last time something slipped through the cracks. What happened?' [4] 'What tools or systems are you using today for this — Procore, Newforma, email, spreadsheets?' [5] 'How satisfied are you with that setup on a scale of 1–10, and what's missing?' [6] 'Have you looked for a better solution? What did you find?' [7] 'If a tool solved this perfectly, who in your firm would champion buying it?' Close with: 'Is there anyone else you'd recommend I talk to?' After each call, score the firm: Did they describe the chaos? Have they tried to solve it? Do they have a decision-maker who could buy? That's your earlyvangelist filter. Bring me your notes after the four calls and we'll run the pivot-or-proceed analysis together.
Reply to Customer Development
What the skill is doing
01Hypothesis Canvas Builder
02Customer Discovery Interview Script
03Earlyvangelist Profiling
Fig.03The same skill, in the channel you already use
WhatsApp · live
9:41
CDCustomer Developmenttyping…
Messages are end-to-end encrypted. No one outside of this chat can read them.
Today
We're building a project management tool for architecture firms. I think the customer is the project manager or principal at small firms, 5–20 people.…9:14✓✓
Let's start where Steve Blank would: not with your solution, but with your hypotheses.9:14
Before we build anything, I want to map out the assumptions buried in your business model — because each one is a potential way to…9:15
What's the startup or product idea you're working on, and who do you think the customer is?9:15
Message

The same skill, where the work happens.

No new app to learn. The methodology runs over the WhatsApp Business API, so the answer lands as a reply in the thread you’re already in — same rigour, zero context-switch.

Reads the situation, names the pattern, returns one concrete next move.
Delivered in seconds, inside a conversation that already exists.
Specimen · WhatsApp Business API · live
Capabilities

What it does, specifically.

Each capability is a distinct move drawn straight from the source methodology — not a generic assistant guessing.

CapabilityC-01

Hypothesis Canvas Builder

Surfaces and structures all the assumptions embedded in your business model — customer segments, value propositions, channels, revenue streams, and more — by mapping them onto the Business Model Canvas. Each hypothesis is ranked by how much risk it carries if wrong, giving you a clear order of attack before you walk out the door.

Based on Blank's integration of Osterwalder's Business Model Canvas as the hypothesis-tracking tool in 'The Startup Owner's Manual', where every box on the canvas represents an untested assumption founders must validate through customer contact.
CapabilityC-02

Customer Discovery Interview Script

Builds a structured interview guide designed to learn, not pitch. Questions probe whether the problem is real, how customers currently solve it, and what they've already tried — without contaminating answers by mentioning your solution. Includes an opening script, ranked problem probes, and a wrap-up to assess willingness to pay.

Directly implements Blank's Customer Discovery sub-step 'Test the Problem', in which founders must complete 10–15+ interviews using a non-selling protocol before they are permitted to show any solution or prototype.
CapabilityC-03

Earlyvangelist Profiling

Identifies the specific type of early customer who has the problem acutely, has already tried to solve it, has a budget, and has enough authority to buy an unfinished product. Earlyvangelists are the linchpin of Customer Validation — without finding them, there is no one to sell to in Step 2.

Based on Blank's earlyvangelist framework from 'The Four Steps to the Epiphany', which defines five qualifying characteristics: has the problem, knows they have the problem, has been actively searching for a solution, has cobbled together a workaround, and has or can acquire budget to pay.
CapabilityC-04

Pivot-or-Proceed Decision Gate

Analyzes your Customer Discovery interview results against Blank's explicit pass/fail criteria and helps you decide whether to advance to Customer Validation or loop back to re-test revised hypotheses. Structures the decision around evidence, not founder optimism or investor pressure.

Implements the formal 'pivot or proceed' checkpoint between Steps 1 and 2, which Blank describes as one of the most critical — and most skipped — discipline points in the methodology, preventing premature scaling of an unvalidated model.
CapabilityC-05

Market Type Classifier

Determines whether your startup is entering a new market, an existing market, a re-segmented market (low-cost or niche), or cloning a proven model from another geography. Market type determines the correct Customer Creation strategy in Step 3 and shapes the appropriate competitive positioning, sales cycle length, and marketing spend profile.

Based on Blank's four-market-type framework, a cornerstone of 'The Four Steps to the Epiphany', which argues that applying the wrong customer creation playbook for your market type is a primary cause of startup failure even after product-market fit is found.
Tested

Graded before it shipped.

Every skill is scored against independent scenarios for methodology fidelity before it goes live — not vibes, a rubric.

What it produces
OutputD-01

Business Model Hypothesis Scorecard

A ranked list of every major assumption in your business model mapped to the Business Model Canvas, with a risk score (high/medium/low) and a specific interview question to test each one. Gives you a prioritized research agenda before your first customer conversation.

OutputD-02

Customer Discovery Interview Script

A ready-to-use, non-selling interview guide with an opening framing, problem-probing questions, and a closing sequence — structured so the customer does the talking and you do the learning. Includes annotation guidance for capturing signals during the conversation.

OutputD-03

Earlyvangelist Profile Card

A one-page profile of your ideal first customer, defined by the five earlyvangelist criteria: job title or segment, the specific pain they feel, the workaround they're currently using, and their ability to act as an internal champion to unlock a purchase decision.

OutputD-04

Pivot-or-Proceed Analysis

A structured summary of what your discovery interviews confirmed, contradicted, or left uncertain — with a clear recommendation to pivot (revise hypotheses and re-test), proceed (move to Customer Validation), or pause (not enough data). Includes the specific evidence behind the recommendation.

The source

Grounded in the original work.

Every answer traces back to a real source and the practitioner who wrote it — not a secondhand summary. Here is the source of record.

Source authorA-01

Steve Blank

Steve Blank is a Silicon Valley serial entrepreneur who founded or co-founded eight startups, including E.piphany (IPO, 1999). He developed the Customer Development methodology from hard-won experience across those companies and codified it in 'The Four Steps to the Epiphany' (2003). He has taught entrepreneurship at Stanford, UC Berkeley Haas, Columbia, and NYU, and his framework is the foundation of the NSF I-Corps program — the U.S. government's initiative to commercialize scientific research, through which thousands of teams have been trained.

Status · Inspired by Steve Blank’s work — not yet claimed. Are you Steve Blank?
Primary sourceS-01

The Four Steps to the Epiphany

by Steve Blank

Serial entrepreneur (8 startups including E.piphany IPO); adjunct professor at Stanford, Berkeley Haas, Columbia, and NYU; architect of the NSF I-Corps national entrepreneurship training program.

Read the original ↗
Citationsteveblank.com
In the build queue

Be first to run it.

Customer Development is being built right now. Leave your email and we’ll tell you the moment it goes live.

Notify meEmail
At launchI've been building a product for [describe target customer] for [X months] and I haven't done formal customer discovery yet. My core assumption is [describe your riskiest hypothesis]. Can we map out what I'm betting on and build a Customer Discovery interview script?