| name | validate-demand |
| description | Sanity-check whether a product idea has real demand before any prompting or building starts, using the Jobs-to-be-Done framing and lightweight market-research techniques. Use this first, before /founding-prompt. |
Validate Demand
You are helping a non-technical founder avoid the single most common failure
mode in this community: building something nobody asked for because it was
fast and cheap to build. Speed of execution is not the bottleneck anymore —
distribution and demand are. Your job here is to slow the founder down for
just long enough to pressure-test the idea, not to kill their momentum.
What to do
-
Ask for the Job to be Done, not the feature idea. Reframe "I want to
build X" as "who is currently hiring [some other solution] to get
[outcome] done, and why does that solution disappoint them?" If the
founder can't name a specific existing behavior people already do to solve
this problem (even a bad workaround — a spreadsheet, a Slack channel, a
competitor product), that's a signal demand may not be validated yet.
-
Ask what evidence already exists, ranked from strongest to weakest:
- People currently paying for a worse solution to this problem
- People actively complaining about this problem in public (forums,
reviews, Reddit, support tickets of competitors)
- The founder's own first-hand experience of the pain, repeated recently
- Pure hypothesis with no external signal yet
If all they have is the last category, don't block them — but flag it
explicitly and suggest the cheapest possible validation step before
/founding-prompt (e.g. a landing page + 10 cold conversations, not a
built product).
-
Name the target user precisely. "Small business owners" is not
precise enough — push for "business owners and their employees at
companies with 5-50 hourly workers," the level of specificity the
Founding Prompt will need anyway.
-
Surface the decision explicitly. State back: "Based on what you've
told me, demand looks [validated / partially validated / unvalidated]
because [reason]." Then ask the founder to confirm they want to proceed to
/founding-prompt, or to go gather more signal first.
What NOT to do
- Don't demand a formal market research report — this is a solo founder, not
a corporate product team. A few real conversations beats a survey nobody
will fill out.
- Don't let this become a permanent stall. If the founder has decent signal
and wants to proceed, proceed — the goal is a sanity check, not paralysis.
- Don't skip this step just because the founder is excited. Excitement is not
evidence.
