The first ten customers who are not doing you a favour

Early demand is easy to manufacture and hard to read. The discipline is separating people who want the problem solved from people who want to be supportive.

Interest expressed in a conversation costs nothing, so it is produced generously. The useful question is not whether someone likes the idea; it is what they have already done about the problem before meeting you. Someone running a painful spreadsheet workaround every week has demonstrated demand. Someone who finds the idea interesting has demonstrated politeness.

Signals, ranked by how expensive they are to fake

• Paid, in advance, without a discount. Hard to fake, hard to obtain.
• Existing workaround in production. They built something ugly because the pain was real.
• Introduced you to a colleague unprompted. Reputation risk is real currency.
• Agreed to a pilot with a named owner and a date. Cheap, but not free.
• "Let us know when it's ready." Not a signal.

Why the first segment should be narrower than comfortable

A narrow segment makes the product legible: the same objections recur, the same integrations matter, and word travels because the people know each other. A broad early segment produces an average of several products and a roadmap that satisfies nobody in particular.