The constraint test (three questions, scored 1-5 each)
Three questions. Each gets a score from 1 to 5. Higher is better. Add the three. Highest total wins.
If two ideas tie, pick the one where the answers came out of you fastest. Speed is a signal that you actually thought it through, instead of constructing the answer just now.
The three questions, named:
- Reach. Can you actually hand the v1 to this person?
- Frequency. How often does the pain hit?
- Buildability. Can you ship v1 in four weekends?
Each gets a section. Each has a tell for when you're cheating yourself.
Question 1: Reach
How fast can you get this person to use a rough v1?
- 5. You live with them or talk to them weekly. You could hand them the v1 by next Saturday.
- 4. You know them but don't see them often. You could ping them and they'd respond within a week.
- 3. You know them through someone else (a friend's older sibling, a parent's colleague's kid). Two weeks to get an intro.
- 2. They exist in a community you're in (a Discord, a subreddit). You'd have to post and hope someone bites.
- 1. They're a hypothetical "type of person." You don't actually know anyone like them yet.
Why this matters more than you think: in Module 6 you need a real human to pay for a real product. The closer you are to that human now, the lower the friction in Module 6. A short walk to your sister's room beats four months "growing an audience" on Instagram.
If your candidate is a 1 or a 2 on Reach, that's your warning. Strong consider picking someone from your Lesson 1 list you can actually reach.
Frequency next.
Question 2: Frequency
How often does your customer hit this problem?
- 5. Every day. Multiple times.
- 4. Every couple of days.
- 3. Once a week.
- 2. Once a month.
- 1. Once a year or less.
Why this matters: people pay to make pain go away. Daily pain is worth far more than annual pain. A daily annoyance at $5 a month is an easier sell than an annual one at $60 a year. Same math. The person FEELS the daily pain ten times before they remember they paid for the annual one.
Refinement: if the pain is technically "once a year" but the person spends three concentrated weeks on it, score it higher. Tax season is once a year, and TurboTax is a $4 billion business. College apps is once a lifetime, and Tara's tracker was a real product. Concentrated pain counts.
The tell you're scoring too high: if you write "well, kind of every few weeks, sort of," you're rationalizing. Once a month is a 2, not a 4. Score for the calm version of you that will read this in November.
One question left, and it's the one where everyone overestimates themselves.
Question 3: Buildability (the four-weekend test)
Imagine the dumbest possible version that just barely makes the pain go away. Not polished. Not pretty. The version with no settings panel, no login, no dashboard, no "user accounts," no "AI-powered" anything. The version where the pain just barely stops.
Could you build THAT version in 4 weekends (8 days of focused work)?
- 5. Yes. You could sketch the entire thing on paper right now and the build is 3 to 5 hours.
- 4. Yes. One weekend of focused work.
- 3. Yes. Two or three weekends.
- 2. Maybe four weekends. There's complexity you can't see yet.
- 1. More than four weekends, and you're handwaving about how to even start.
Why this matters: Wright is 12 modules. Modules 3 and 4 are where you build. If you can't ship something rough in those modules, you can't ship anything later. The whole point is to put a real thing in front of a real human and learn what happens. You can't do that with a v3.
The tell that you're rating too high: if your "v1" includes the words "login," "user accounts," "database," "admin panel," "social feed," "gamification," or "AI-powered," you're not imagining a v1. You're imagining a v3. Strip it.
Vygotsky has a concept called the zone of proximal development, by the way. It's the band of things you can't do alone but you can do with help. Your v1 lives in that band. Too easy and it's not worth shipping. Too hard and you stall out. Just above what you can do alone, with the AI builders as help: that's the v1. (More on this in Module 3.)
Time for the math.