Month 10-Week 3: Capstone build sprint 2 + first user¶

Week summary¶

Goal: Continue building. Find your first external user. Watch them use the tool unaided for 15 minutes. Fix the top 3 confusions.
Time: ~9 h over 3 sessions.
Output: Feature-complete v0.1 capstone; user feedback; top 3 confusions addressed.

Why this week matters¶

The 15 minutes of watching an unaided user is the highest-leverage 15 minutes of the entire month. It surfaces invisible-to-you problems: unclear README, broken installs, opaque error messages.

Prerequisites¶

M10-W02 complete.

Recommended cadence¶

Session A-Tue/Wed evening (~3 h): finish core features
Session B-Sat morning (~3.5 h): find user + observe
Session C-Sun afternoon (~2.5 h): fix top confusions + push

Session A-Core feature freeze¶

Goal: Hit "feature-complete v0.1"-the minimum viable capstone.

Part 1-Triage (15 min)¶

What 2-3 features remain to call it v0.1? Pick. Cut everything else.

Part 2-Build (150 min)¶

Part 3-Push v0.1 release candidate (15 min)¶

Tag v0.1.0-rc.1.

Output of Session A¶

Capstone at v0.1-rc.

Session B-Find user + observe¶

Goal: One real user attempts to use your tool unaided. You watch. You don't help.

Part 1-Find users (45 min)¶

Reach out to 3-5 people. Targets: - A coworker who knows nothing about your specialty. - A Discord acquaintance in the field. - A Twitter/X contact. - A friend who codes.

The ask:

"I'm building [X]. Would you spend 15 minutes trying to install and run it, telling me what's confusing? I won't help during-your unfiltered confusion is the gold."

Two will say yes.

Part 2-Observe one user (45 min-synchronous if possible)¶

Schedule 30 min on a call with screen-share. They drive.

Watch them: - Open the README. - Try to install. - Try to run an example. - Try to do their own task.

Write down every confusion. Don't help unless they're truly stuck.

Part 3-Triage (60 min + post-call)¶

Likely top confusions: - README jumps too fast. - Install instructions are missing a step. - Error messages don't tell them what went wrong. - Example doesn't show what they actually want to do.

Pick the top 3 to fix.

Output of Session B¶

User session recorded (with permission) or notes.
Top 3 confusions identified.

Session C-Fix top 3 + retro¶

Goal: Address the top confusions. Push v0.1.0.

Part 1-Fix README + install (45 min)¶

Most user confusion is about onboarding. Fix: - Quickstart 3-5 commands. - Common errors documented with fixes. - An example that mirrors a user's likely task.

Part 2-Fix the deepest issue (45 min)¶

If the user got stuck on a core flow, fix the flow OR add a "common pitfalls" doc page.

Part 3-Push v0.1.0 (45 min)¶

Tag. CHANGELOG. Brief release-notes post on your blog (if you have one).

Output of Session C¶

v0.1.0 tagged.
Top 3 confusions fixed.

End-of-week artifact¶

v0.1.0 tagged with feature freeze
One real user observed
Top 3 confusions fixed

End-of-week self-assessment¶

My capstone has been used by a stranger.
I know what was confusing.
I've fixed the most painful onboarding issues.

Common failure modes for this week¶

No user. "I'll find one later"-usually doesn't happen.
Helping during the observation. Don't. Their confusion is data.
Defending the design when criticized. Listen first.

What's next (preview of M10-W04)¶

v0.1 ship: tests + CI green + README excellent + soft launch.