06-reference

commoncog how moats are built

2026-07-21·reference·source: Commoncog·by Cedric Chin

"How Moats Are Built" — @CedricChin

Why this is in the vault

Explains why sustainable competitive advantage is easy to identify retroactively but requires effectuation — trial-and-error under uncertainty, hidden from competitors — to actually build, which is the exact playbook RDCO is running.

The core argument

Hamilton Helmer's 7 Powers framework makes moats legible post-hoc: you can look at a successful business and cleanly map it to Switching Costs, Scale Economies, Network Effects, and so on. The issue Cedric addresses is that this diagnostic clarity creates a false impression that moat-building is similarly legible in real time.

It isn't. The piece argues that building a Power requires two conditions: operating outside your competition's field of vision (hiding while you develop the capability), and running genuine experiments under uncertainty rather than executing a pre-formed plan. This is the effectuation frame — act first, learn from the environment, refine — not the causal/predictive model of strategy.

The theoretical backbone points to two prior Commoncog pieces: "When Action Beats Prediction" and "How to Run Smart Experiments When You Just Don't Know." Moat-building, on this account, is a downstream output of running many smart experiments while your competitors are focused elsewhere.

Mapping against Ray Data Co

The phData DSA bet is a live case of this thesis. Ben is building a Data+AI consulting expertise moat by operating inside a firm (phData) that gives him deal flow and client exposure while the work remains largely invisible externally — the "hide from competition" condition is satisfied by default. The moat compounds through cert escalators (Snowflake GenAI + Anthropic Claude Certified Architect) that lock in credentialed differentiation before the market saturates.

The RDCO AI COO deployment is the effectuation layer: rather than planning the perfect agentic system, Ray is deployed now and iterated against real operating conditions — unhobbling is not a pre-formed roadmap, it's a series of experiments (plugin topology, skill installs, cron agents, channel comms) run under genuine uncertainty. The moat being built is operational AI fluency + tooling depth that won't be visible to competitors until it's structurally durable.

The "Fourth Career Moat Pattern" note already in the vault makes the individual dimension explicit. This piece adds the mechanism: moats aren't planned, they're grown by effectuation.

Issue contents

Related