1. Make It Legible
A decision isn't real until it's written down:
- what we chose,
- why we chose it,
- who owns it,
- when we revisit it.
Legibility prevents drift.
It also kills two common failures: ghost decisions and false alignment.
2. Prefer Reversible Moves with Learning Value
Most decisions aren't existential.
I bias toward low-blast-radius experiments that produce signal quickly—and decisions that produce human learning.
If the cost of being wrong is low and the cost of delay is high, I ship and learn.
3. Protect the Critical Path
Not everything is equally important.
Flow matters more than activity.
I invest my attention where blockage creates downstream drag—across teams, interfaces, handoffs, and decisions that stall momentum.
4. Hold a High Bar—Define "Good" Concretely
A high standard without criteria is just pressure.
A high standard with examples, tests, and boundaries is leadership.
I avoid abstraction. I define "good" so the team can hit it without guessing.
5. Push Decisions to the Edge
The people closest to the problem should make the call—if guardrails are explicit.
My job is to set constraints, not hoard judgment.
Decentralized decision-making is only safe when the boundaries are crisp.
6. Share Reasoning, Not Just Conclusions
A conclusion teaches nothing.
Reasoning creates judgment.
I show how I weigh risk, sequence work, and trade off constraints so others can generalize it into domains I'm not in.
7. Optimize for Momentum Over Elegance
Elegance is a luxury.
Momentum compounds.
When choosing between perfect and shipped, I bias toward shipped—as long as the deviation from standards is explicit and deliberate.
8. Expose Risks Early
Risks don't get better with age.
I raise them at 20% confidence, not 80%.
Early signal buys optionality; late signal buys firefighting.