More people. More tools. More budget. More automation.
It feels like progress. It's often the opposite.
Here's the clearest proof I know of.
By mid-2018, Tesla's Model 3 line had a first pass yield of around 14%. Industry average is 65-80%. Roughly 86 of every 100 cars needed rework before they could ship. Cash was burning, customers were waiting, and Wall Street was circling.
The plan had been the "alien dreadnought": a fully automated factory of robots and conveyors that would outperform humans. The robots became the bottleneck.
So Tesla did the opposite of "add more." In June 2018, they built a general assembly line in a tent outside the Fremont factory. It was manual labour. They called it GA4. It went up in three weeks and lifted Model 3 output by about 50%.
Within weeks, Tesla hit 5,000 cars a week.
Musk's own diagnosis: "Excessive automation at Tesla was a mistake. Humans are underrated."
This is Goldratt's Theory of Constraints (The Goal, 1984). Every system has exactly one constraint that limits everything else. Five steps to fix it:
- Identify the real constraint (not the loudest problem).
- Exploit it. Get every drop from it before adding anything.
- Subordinate. Make every other process serve it.
- Elevate. Add capacity only after steps 1-3.
- Repeat. The constraint will move.
Most leaders jump straight from step 1 to step 4. They spot the bottleneck and immediately throw resources at it. Tesla did exactly that with automation, and the robots became the new constraint.
The answer is almost never more. It's almost always simpler.
If you lead an ops, product, or engineering team, try this before your next hire or tool purchase: name your single biggest constraint, then ask whether you've truly exhausted steps 2 and 3 before spending on step 4.
What's the one bottleneck your team keeps building around instead of confronting? Comment below in one word.
