Part II · Find Value — Chapter 10
Listen before you build
Customer discovery is not asking people to praise your idea. It is reconstructing what they already do, why it fails, what it costs, and how decisions are made.
Speak to at least 30 relevant people before committing substantial time or capital. The number is not magical. Its purpose is to prevent one enthusiastic conversation from becoming a false market.
Ask about the past, not imagined behaviour
Bad question:
Would you use an AI tool that saves time?
Better questions:
- • Walk me through the last time this happened.
- • What triggered the work?
- • Who handled each step?
- • Where did it slow down or fail?
- • What did the failure cost—in time, money, customer trust, or risk?
- • What have you tried already?
- • What do you pay today, including employee time and other tools?
- • Who owns the budget and who can veto a change?
- • What would have to be true for you to switch?
- • What data, security, integration, or compliance concerns would stop you?
- • If this remains unsolved for a year, what happens?
Do not pitch during the first part of the conversation. Silence is useful. Ask for artifacts where appropriate and permitted: blank forms, screenshots with sensitive data removed, workflow diagrams, reports, or sample outputs.
Capture evidence consistently
For each conversation, record:
| Field | What to capture |
|---|---|
| Segment | Company type, size, role, relevant tools |
| Trigger | What starts the workflow or problem |
| Current method | Steps, people, systems, and hand-offs |
| Field | What to capture |
|---|---|
| Pain | Specific delay, error, cost, anxiety, or lost opportunity |
| Existing spend | Vendor fees and internal labour |
| Decision process | Buyer, approvers, blockers, timing |
| Exact language | The words the customer uses to describe the problem |
| Commitment | Introduction, data sample, next meeting, pilot, or payment |
After each small, capacity-appropriate batch of interviews, look for patterns. Do not average away different segments. A problem for a 10-person agency may have a different buyer and workflow in a 1,000-person enterprise.
The commitment ladder
Interest is cheap. Ask for progressively stronger evidence: 1. time for a second conversation;
2. introduction to the process owner or buyer; 3. access to a sanitized example or workflow; 4. agreement on success criteria; 5. signed pilot scope; 6. payment or deposit; 7. active usage;
8. renewal, expansion, or referral.
A small number of people who pay teach more about a business than a large audience that merely clicks “like.”
RED LINE Do not extract confidential data, record calls without permission, or paste customer information into third-party AI tools without authorization and appropriate safeguards. Trust lost during discovery cannot be repaired by a good demo.
· GE ch. 9; TABLE pp. 13–16; SALES pp. 20–21. Mechanisms paraphrased; judgment and examples are labelled.