Plan Heavy vs Ship Light

Paul Graham once wrote about startups being default alive or default dead — a brutally simple way to check whether a company’s current trajectory leads to survival or collapse.

I like the concept of defaults: what are the organizational tendencies that arise as a result of the cultural revealed preferences? One such dichotomy is the need to build heavy plans, vs the impulse to ship light.

---

At larger companies, the default is to Plan Heavy.

• To ensure success, the instinct is to make something bigger — more features, more teams, more cross-functional alignment.
• Bigger means the stakes rise, as does the cost of being wrong.
• Stakes drive prep work and process.
• Prep work drives expense.
• Which, ironically, makes it harder to take risks or ship fast.

Planning becomes the product. Execution is gated by alignment. The illusion of control replaces actual learning.

Yet there is another dynamic at play: large organizations often say they want to be both fast and predictable.

In practice, those goals fight each other.

Improving speed of delivery and predictability of delivery at the exact same time is extremely difficult, especially if you also claim you want to be nimble and responsive to market feedback.

Software engineers, in particular, struggle to accurately estimate work when:

1. The work is not well understood. This happens when it is being done for the first time, or when it is scoped far in advance. Good product requirements take time to develop. Given limited time, product managers will prioritize the PRDs for work happening next over work happening later. The “later” work is, by definition, less defined.
2. The work is not happening soon. The person estimating it may not be the person doing it. The codebase may change. The context and rationale may fade. When it finally starts, someone has to page the problem back into memory. That long context switch alone destroys estimation accuracy.

The focus on predictability results in an exercise in estimating the unknown. As a result, the natural response is defensive estimation. Larger buffers. Sandbagging. Bigger milestones. Heavier plans.

Predictability increases.
Actual velocity and reactivity to market feedback decreases.

---

At small companies, the default is to Ship Light.

• To ensure survival, the instinct is to make something smaller — smaller POCs, fewer people, tighter feedback loops.
• The goal is to minimize waste on things that might not work.
• The cheaper the failure, the faster the learning.
• This keeps everyone closer to the truth of what customers actually want.

High-throughput organizations do not attempt to predict the distant future with precision.

Instead, they shorten the feedback loop.

They operate on iteration cycles small enough that estimates are informed by fresh context, stable teams, and real feedback.

Long-term roadmaps become directional.
Commitments skew near-term, and carry more weight.
The rest is best effort, not false precision.

---

The real tension

Highly predictable roadmaps usually require heavier plans and larger estimates.

Highly adaptive organizations accept that long-range precision is a guess, and focus instead on rapid iteration.

You can optimize for certainty, or you can optimize for learning velocity. It is very hard to maximize both at once.

The real leadership challenge is not choosing one permanently: it is knowing when to Plan Heavy versus when to Ship Light and being honest about which default your organization is currently reinforcing.