You shipped your MVP. It works. It's not perfect—and that's exactly right. Now comes the part that separates good builders from great ones: iteration. Taking what you learned and making it better.
Iteration Tracker
Simple template for capturing feedback and planning improvements
📄 Preview PDF
🔄 The Iteration Cycle
After shipping, you enter a new loop:
Iteration Process
⚡ Pro Tip: Use Before Improving
Resist the urge to jump immediately to Phase 2. Use Phase 1 for at least a few days. You'll discover what really matters vs. what you thought would matter.
📋 Gathering Feedback
The best feedback comes from using your own creation. As you use it, ask:
- What do I reach for that isn't there?
- What's more complicated than it needs to be?
- What breaks or behaves unexpectedly?
- What would make this 10% better right now?
Keep a simple list. It'll become your roadmap for future phases.
🎯 Feedback Capture Template
Things that don't work as expected or cause errors
Example: "Timer resets when I click outside the window"
Things that work but are inconvenient or tedious
Example: "Have to retype project name every time I start timer"
Features you find yourself wishing existed
Example: "Want to see history of what I timed yesterday"
Improvements that would be cool but aren't essential
Example: "Different color themes would be neat"
🎯 Prioritizing Changes
Not all feedback is equal. Prioritize:
Bugs that break functionality. Missing features that make it unusable. Things that are actively annoying every time you use it.
Improvements that would make it nicer. Features from your "Phase 2" list. Things that are inconvenient but workable.
"Nice to have" features. Polish and aesthetics. Things that don't affect core functionality.
✨ The Compound Effect in Action
Each iteration session starts faster because Claude.md remembers everything. Your documentation from the "Compound" step pays off every single time you return to the project.
📖 Complete Iteration Example
Let's watch how the Learning Activity Timer evolves through iteration:
Phase 1 Shipped: Basic start/stop timer with elapsed time display
User uses the timer for 5 work sessions.
Observations:
• Works great for timing one thing
• Annoying: Can't label what I'm timing
• Missing: No record of past tasks
• Bug: Time doesn't persist if I accidentally refresh
High Priority (Phase 2):
1. Add task name input
2. Save completed tasks to history
3. Fix: Persist current timer state
Medium Priority (Phase 3 maybe):
• Visual time breakdown
• Edit/delete past tasks
Low Priority (later):
• Export to CSV
• Statistics dashboard
Opens Claude Code in the project folder.
Prompt: "Claude, read the Claude.md file. I want to add three features for Phase 2: [lists features]. What's the plan?"
Claude immediately understands the project (thanks to Claude.md!), proposes approach, builds features in 45 minutes.
Result: Can now name tasks, see history, timer survives refresh
Uses improved version daily.
New observation: "I want to see total time by task across days. Like 'I spent 6 hours on Module 3 this week.'"
Adds to Phase 3 list. But not urgent—Phase 2 is already super useful.
✅ Success Pattern
Notice the cycle: ship → use → learn → prioritize → build → repeat. Each phase makes it more useful. But each phase is independently valuable.
🚀 From MVP to Real Product
Through iteration, your MVP evolves:
- Phase 1: It works (MVP)
- Phase 2: It works reliably (stability)
- Phase 3: It works well (features)
- Phase 4: It looks good (polish)
- Phase 5+: It delights (refinement)
Each phase builds on the last. That's compound engineering.
💡 When to Stop Iterating
How do you know when a project is "done"? Three signals:
- It solves the original problem: The pain point you started with is gone
- You're using it regularly: It's become part of your workflow
- Improvements feel optional: New features would be nice, not necessary
At that point, you can:
- Maintain: Keep using it, fix bugs if they appear
- Polish: Make it prettier, add those "nice to have" features
- Share: Show it to colleagues, maybe they'd find it useful
- Move on: Start a new project, apply what you learned
⚡ Pro Tip: One Project at a Time
It's tempting to start three projects after finishing this course. Resist. Build one thing all the way through Phases 1-3. Learn the full cycle. Then start the next one.
🎓 Course Complete!
You've learned:
- ✅ Terminal basics (Module 0)
- ✅ How to use Claude Code (Module 1)
- ✅ The planning mindset (Module 2)
- ✅ The compound engineering loop (Module 3)
- ✅ How to scope, phase, ship, and iterate (Module 4)
You're not a vibe-coder anymore. You're a builder.
🚀 What's Next?
You have everything you need to build AI-powered tools that make your work easier. Here's your roadmap:
Your Next Steps
🎯 Your Advantage
As an instructional designer, you already know how to break down complex processes, write clear instructions, and design for the user. These skills give you a huge advantage in building with AI. You're not learning to code—you're learning to instruct. That's what you do best.
📚 Resources & Further Learning
- Every.to Chain of Thought Dan Shipper's ongoing exploration of AI and productivity
- Anthropic Documentation Official docs for Claude and Claude Code
- Claude Code in Action (Anthropic Academy) Free course on advanced Claude Code features
- Compound Engineering Plugin Open-source tools from Every for compound engineering
💭 Final Reflection
Before you close this course, take a moment to consider:
- What's different about how you think about AI tools now vs. when you started?
- What project are you most excited to build?
- What surprised you most in this learning journey?
- How will you stay accountable to actually building (not just learning)?
🎉 Congratulations!
You've completed Claude Code Fundamentals. Now go build something amazing.
Course Complete • 20 Topics • Ready to Build