FEATURES NEVER BUILT: 0
THE GROWING TEAM
A Systems Thinking Simulation
Press SPACE to begin
Developers
QA Engineers
BUILD
TEST
SHIP
10 built/week → 5 tested/week → 5 shipped
Developers: 10 QA: 2 Backlog: 0
5
features / week
Budget for 5 new hires.
Where do you invest?
👨‍💻
+5 Developers
More builders = more features, right?
15 Dev / 2 QA
🔍
+5 QA Engineers
Make sure quality keeps up with growth.
10 Dev / 7 QA
⚖️
Split 2 + 3
More QA than devs. Balance the pipeline.
12 Dev / 5 QA
What happens with each strategy?
+5 Developers
BUILD15
TEST5
5
UNCHANGED
+5 hires → ZERO improvement
Split 2 + 3
BUILD12
TEST12.5
12
+140% BEST
Balances the pipeline
+5 QA Engineers
BUILD10
TEST17.5
10
+100%
Addresses the constraint
Any strategy that addresses the constraint works. Balancing the pipeline works best.
OK, let's address the constraint.
10 devs + 5 QA. Output: 10/week. Now let's see what happens over time.
10
features / week
The constraint is cleared. Output doubled.
But what happens next?
Before we simulate — a quick test.
Your team ships 10 features/week.
Each shipped feature needs 1 hour/week of maintenance — forever.
Your 10 devs have 400 hours/week total.
At Week 30, what will output be?
Type a number in chat — or just hold it in your mind.
0
2
4
6
8
10
Week 0
Output: 10.0 / week
PAUSED — check your prediction
■ New Features ■ Maintenance
Maintenance: 0%
Before I show you what happens next —
You just watched output drop from 10 to 4.1.
The obvious response: hire 5 more developers.

Think about what you've just seen.
Will hiring more help, hurt, or make no difference?
Think about it. If you can explain WHY — type your reasoning in chat.
Output is dropping.
What do you do?
📈
Hire More Developers
Compensate for lost capacity.
Add resources to maintain velocity.
🛠️
Slow Down & Fix
Reduce technical debt.
Deprecate unused features.
Developer
Headcount
Feature
Output
Shipped
Features
Maintenance
Burden
Available Dev
Capacity
(+)
(+)
(+)
(-)
(-)
REINFORCING
LOOP
Output is dropping again.
How fast do you react?
⚡
React FAST
Check weekly. Hire immediately
when output dips below target.
Hire 5 devs per batch
🐢
React SLOW
Check quarterly. Hire gradually.
Trust the system to stabilize.
Hire 1-2 devs per batch
Two strategies. Same system. Same starting point.
Strategy A: React FAST — check weekly, hire 5 devs whenever output dips.
Strategy B: React SLOW — check quarterly, hire 1-2 devs gradually.
Which produces more output after 1 year?
Type FAST or SLOW in chat — or decide in your head.
⚡ REACT FAST
🐢 REACT SLOW
OSCILLATION
Wild swings. Team bloats to 20+. Unpredictable.
CONVERGENCE
Steady growth. Team stays lean. Predictable.
⚡ React FAST — why it oscillates
Week 0: Output drops. You hire 5 devs.
Weeks 1-8: Nobody arrives yet. Hiring delay = 8 weeks.
Week 4: Still dropping. You hire 5 MORE. Can't see the pipeline.
Week 8: First batch arrives. Mentoring eats senior devs' time.
Week 12: Second batch arrives. Team = 20. Overshoot.
Result: You're always reacting to OLD data.
🐢 React SLOW — why it converges
Quarter 1: Output drops. You hire 2 devs.
Weeks 1-8: You wait. You trust the delay.
Week 8: 2 arrive. Small mentoring burden.
Quarter 2: Reassess with CURRENT data. Adjust gradually.
Week 26: Team stabilizes. Output steady.
Result: Response speed matches system delays.
Fast reaction: 20+ devs for unpredictable output. Slow reaction: 12 devs for steady output.
The system rewards patience, not speed.
#1
🔍
THE CONSTRAINT
A system can only move as fast as its slowest part. Improvement elsewhere is an illusion.
#2
🔄
FEEDBACK LOOPS
Your success feeds your failure. More shipping creates more maintenance, which kills your capacity to ship.
#3
⏳
DELAYS
In systems with delays, reacting faster means oscillating harder. Speed without understanding is chaos.
CONSTRAINT → LOOP → DELAY
One more scenario. Can you diagnose it?
A fintech deploys an AI chatbot for customer support.
Resolution time drops from 4 hours to 45 minutes.
Customer satisfaction RISES for 3 months.
Month 4: satisfaction CRASHES — resolution time is still 45 minutes.

What happened? Use all three lenses.
Hint: this involves a constraint, a feedback loop, AND a delay. Type your diagnosis in chat.
Right now — in your company —
where is the orange eating the cyan?
Your team. Your product. Your pipeline.
Where is what you've already built consuming your capacity to build more?
If you can see it — you've already started thinking in systems.
■ Capacity to build new features ■ Maintenance of what you've already shipped