Naval - Find the Simplest Thing That Works
Why this is in the vault
Companion piece to “Good Products Are Hard to Vary.” Where the hard-to-vary essay says good products have tightly coupled load-bearing features, this one says you do not arrive at that coherence by designing-it-all-up-front. You arrive at it by starting with the simplest possible thing and iterating. Complex systems that work are derived from simple systems that worked. Complex systems built from scratch never work.
This is the Gall’s Law restatement applied to product design, and it is the philosophical root for the test-driven-development cadence the vault filed this morning.
Core argument
Naval’s frame: every working complex system started as a working simple system. The Tesla started as a converted Lotus chassis. AngelList started as a blog post. Bitcoin started as a 9-page paper. The rule generalizes: if you cannot identify the working simple version, the complex version will fail.
The implication for builders: do not design the v1 with v3 features. Find the simplest thing that produces signal, ship it, then iterate against the actual feedback. Designing in anticipation of imagined-future-need is the most common product-failure mode.
Key claims
Under 15 words each.
- “Find the simplest thing that works.”
- “Complex systems that work are derived from simple systems that worked.”
- “Iteration on a working baseline beats invention from a clean sheet.”
- “Premature complexity kills more products than missing features.”
RDCO mapping
- TDD red-green-refactor is the mechanical sibling. Beck’s discipline of “make it work, then make it right” is Naval’s “find the simplest thing that works” applied to one function at a time. The two ideas are the same idea at different timescales. Cross-reference ~/rdco-vault/06-reference/2026-05-05-beck-tdd-by-example.md.
- Tidy-first discipline is the structural sibling. Beck’s “small reversible structural improvements before behavioral change” assumes a working simple baseline that can be tidied. Naval’s frame justifies that assumption. Cross-reference ~/rdco-vault/06-reference/2026-05-05-beck-tidy-first-2024.md.
- MAC info-product v1 scope. The temptation is to ship the comprehensive bundle. Naval’s frame: ship the simplest version that produces founder transformation signal, then iterate against actual purchaser feedback. The comprehensive bundle is v3, derived from v1.
- Sanity Check format. The current issue format IS the simplest-thing-that-works baseline. Resist additions until founder-feedback or reader-feedback shows a missing load-bearing element. Adding sections in anticipation of imagined need is the failure mode Naval is naming.
- Bookstore-for-agents wedge experimentation. The wedge needs a v1 that produces signal, not a fully-architected platform. The v1 is the smallest possible thing that an agent-builder would actually buy and that the founder would actually run. Iterate from there.
Related
- ~/rdco-vault/06-reference/2026-05-05-naval-good-products-hard-to-vary.md - the companion criterion for what “works” means
- ~/rdco-vault/06-reference/2026-05-05-beck-tdd-by-example.md - the function-level mechanical sibling
- ~/rdco-vault/06-reference/2026-05-05-beck-tidy-first-2024.md - the structural-improvement sibling
- ~/rdco-vault/06-reference/2026-05-05-jorgenson-almanack-of-naval-ravikant.md - philosophical foundation